virtual reality simulations and - göteborgs · pdf filestudies to elucidate the potential...
TRANSCRIPT
VIRTUAL REALITY SIMULATIONS AND
INTERVENTIONAL RADIOLOGY
MAX BERRY
DEPARTMENT OF RADIOLOGY
INSTITUTE OF CLINICAL SCIENCES
THE SAHLGRENSKA ACADEMY AT GÖTEBORG UNIVERSITY
GÖTEBORG, SWEDEN
2007
3
ABSTRACT
INTRODUCTION: Use of virtual reality (VR) simulators in endovascular interventional
education has become increasingly popular yet many questions surrounding this nascent
technology remain unanswered. While progress has been made in other disciplines such as
endoscopy and minimally invasive surgery, scientific evidence investigating endovascular
simulations remains limited. The general aim of this dissertation was to conduct validation
studies to elucidate the potential for skills acquisition and assessment outside of the
catheterization laboratory using VR simulation. Endovascular skills transfer from VR-Lab to
the porcine laboratory (P-Lab) was also investigated. An economic analysis was performed to
assist in the establishment of a realistic VR implementation strategy. MATERIALS AND
METHODS: Simulator validations were conducted by comparing performance metrics
collected from novices and experienced physicians using Student’s t-test. Performance
metrics were recorded by the simulator while participants treated simulated patients suffering
from renal artery stenosis (RAS) and carotid artery stenosis (CAS). Endovascular skills
transfer was tested using the P-Lab as an approximation of the human catheterization
laboratory. A group of endovascular novices were evaluated in the P-Lab and the VR-Lab
using an objective skills assessment of technical skills (OSATS), yielding a Total Score.
Participants were then randomized into different training groups, put through their assigned
training schema and subsequently re-evaluated in both laboratories. ANCOVA analysis was
conducted to compare the cumulative effect each type of training had on Total Score.
Consumable and rental fees from the skills transfer study were used to calculate the
comparison data for the economical analysis. RESULTS: Face validity was demonstrated for
both the renal and carotid artery stenosis modules. Neither construct validity study produced
results which differentiated between the expert and novice performance metrics except for
fluoroscopic and procedural times. VR-Lab training sessions generated skills which improved
P-Lab performances. VR-Lab training cost less than the P-Lab using our economical analysis.
CONCLUSIONS: Despite demonstrating face validity, VR-Lab simulations should not be used
alone for skills assessment outside of the catheterization laboratory in its present form. Skills
learned in virtual reality transfer favorably to the P-Lab and simulation training seems to
offer a viable alternative of non-clinical training. The VR-Lab affords a more economical
method to teach and practice endovascular skills compared to the P-lab. Further research is
needed to elucidate the relative efficacies of both training methods. 5
ORIGINAL PAPERS
This dissertation is based upon the following articles:
I. Berry M, Lystig T, Reznick RK and Lönn L. Assessment of a Virtual Interventional
Simulation Trainer. Journal of Endovascular Therapy, Apr 2006; 13(2), 237-43.*
II. Berry M, Lystig T, Beard J, Klingenstierna H, Reznick RK and Lönn L. Porcine
Transfer Study: Virtual reality simulator training compared to porcine training in
endovascular novices. Cardiovascular and Interventional Radiology, May-June 2007;
30(3), In press.**
III. Berry M, Hellström M, Göthlin J, Reznick RK and Lönn L. Endovascular Training
using Animals or Virtual Reality Systems: An Economic Analysis. Submitted.
IV. Berry M, Lystig T, Reznick RK and Lönn L. The Use of Virtual Reality for Training
Carotid Artery Stenting: A Construct Validation Study. Submitted.
* Reprinted with kind permission from the Journal of Endovascular Therapy ©International Society of
Endovascular Specialists
** Reprinted with kind permission from Cardiovascular and Interventional Radiology published by Springer
Science and Business Media.
7
CONTENTS
ABSTRACT.............................................................................................5
ORIGINAL PAPERS................................................................................7
BACKGROUND.....................................................................................11
ENDOVASCULAR INTERVENTION .........................................................................11
LEARNING THEORY ............................................................................................12
MOTOR SKILL DEVELOPMENT............................................................................13
ENDOVASCULAR SKILLS .....................................................................................14
Skills Assessment ...........................................................................................................15
Skills Transfer ...............................................................................................................16
Cost Effectiveness ..........................................................................................................16
Ethics..............................................................................................................................17
SIMULATOR VALIDITY ........................................................................................18
CATCH 22 .........................................................................................................20
AIMS ...................................................................................................21
METHODS AND MATERIAL .................................................................23
SUBJECTS..........................................................................................................23
VIRTUAL REALITY LABORATORY .........................................................................25
PORCINE LABORATORY ......................................................................................26
LOGISTICS.........................................................................................................27
Studies I & IV ................................................................................................................27
Study II...........................................................................................................................27
9
Study III .........................................................................................................................28
ENDPOINTS .......................................................................................................29
Objective Endpoints.......................................................................................................29
Subjective Endpoints .....................................................................................................32
STATISTICAL ANALYSIS .......................................................................................32
ETHICAL APPROVAL............................................................................................33
RESULTS.............................................................................................35
PAPER I ............................................................................................................35
PAPER II ...........................................................................................................36
PAPER III..........................................................................................................39
PAPER IV ..........................................................................................................43
DISCUSSION........................................................................................47
SKILLS ASSESSMENT ...........................................................................................47
SKILLS TRANSFER ..............................................................................................50
SUBJECTIVE EVALUATIONS OF THE VR-LAB........................................................51
ECONOMICS ......................................................................................................52
ETHICS .............................................................................................................55
STUDY LIMITATIONS ..........................................................................59
FUTURE RESEARCH ...........................................................................61
CONCLUSIONS ....................................................................................63
SVENSK SAMMANFATTNING...............................................................65
REFERENCES......................................................................................73
10
BACKGROUND
ENDOVASCULAR INTERVENTION
Endovascular intervention is a highly specialized and potentially dangerous procedure
performed by interventional radiologists and other specialists. Currently available treatment
options include percutaneous transluminal angioplasty, stenting, thrombolysis, embolization,
intravascular filters, plaque excision and foreign body removal. Traditionally, basic
catheterization skills were acquired by radiology residents—as well as cardiology and
vascular surgical fellows—while performing routine angiograms in the catheterization
laboratory (Cath-Lab) under the close supervision of an experienced mentor. The practical
skills gained during these diagnostic procedures were directly transferable to the execution of
endovascular interventions. The traditional minimal number of digital subtraction
angiographies (DSA) needed prior to advancing on to endovascular interventions ranged
from 50-100 although the actual number varied by location, sub-specialty and procedure.1-5
However, improved diagnostic imaging modalities such as computed tomography- (CTA) or
magnetic resonance- (MRA) angiography combined with a more critical view regarding the
acceptability of training on patients has led to a decrease in the amount of angiograms,
constituting an ethical dilemma.6-9 Both factors have jeopardized the normal training arena
for interventional fellows.
In response, many institutions have turned to non-clinical training methods such as
animal models, typically adult swine in a pig laboratory (P-Lab), to assist endovascular
fellows in making the jump from routine angiography to interventional techniques.10-14 Such
a strategy acknowledges the technical difficulties and the early learning curves experienced
during transitioning.2,3,15,16 The goal is to provide the trainee a place to learn and practice
which simulates the actual catheterization laboratory while removing unnecessary patient
risk.17,18 Excluding ethics and availability of animals for now, one can confidently state that
healthy pigs are not an accurate simulation of pathologic human anatomy. Nonetheless, it is
in vivo and allows the use of real instruments to practice for safe interventions.
Another solution to these conflicting demands is the ex vivo use of the virtual reality
laboratory (VR-Lab). The rapid advances in computer technology over the last decade has
heralded phenomenal technology such as full procedure virtual reality simulators capable of
reproducing operative experiences outside of the catheterization laboratory.19 VR simulators
11
can systematically expose the trainee to a broad variety of pathologies in a safe, reproducible
environment thereby eliminating the usual randomness of training.20-22 Other unique
advantages available with VR technology, such as objective feedback and the removal of the
need for continual direct supervision have already been identified by other research.20,23
However, in their current forms, endovascular simulators require introduction to the machine
itself and close supervision during training to obtain the maximum benefit and prevent the
development of dangerous habits. These systems allow endovascular fellows to develop basic
psychomotor and procedural skills and fully acquaint themselves with the clinical
environment prior to traditional patient based mentoring. Some proactive specialties have
recommended evaluation, piloting and introduction of VR training into their specialty
training to take advantage of VR’s unique capabilities.14,24,25 Indeed, the surgical fields of
endoscopy and laparoscopy have seen an explosion of research examining the usefulness and
validity of VR simulations which is too numerous to list.
LEARNING THEORY
Cognitive learning—defined as mental couplings between knowledge and physical
skills—states that learning occurs when humans attain equilibrium between their reactions
and their surroundings through assimilation and accommodation.26 Assimilation, according to
Piaget, is simply learning by addition. In other words, as one is exposed to new stimuli, a
repertoire of experiences is added to the existing knowledge base. Accommodation occurs
when new situations are encountered which require the use of pre-existing knowledge. This
process, although not directly creating new knowledge, entails the deconstruction and
rebuilding of existing knowledge into a form which is applicable to the alien situation.
Accommodation presents more difficulty for the learner yet offers a deeper understanding.
Nissen expanded this line of thinking with his notion of cumulative learning, which he
defined as the use of all of ones cognitive schemas applied to a completely new learning
environment—creating a new set of knowledge.27
Psychodynamic learning, as proposed by Vroom, tells us that all learning is dependent
on feelings and motivations.28 Thus the feelings, both positive and negative, we experience
during a learning situation are an integral part of the process and affect the outcome.
Motivation to learn falls within this category as well. A learner who believes that they can
master a new topic which will prove useful in a fashion that improves their own situation
stands a much higher chance of succeeding.27
12
Bruner’s societal learning dimensions focus on the fact that everything a person does
is influenced by, and in turn influences, a cultural and social context.27 Societal learning is
thus an interaction between people, language and objects and is dependent upon social and
contextual influences. A more concrete example of this type of learning is the mentor-
apprentice model.29 The apprentice begins his studies as a passive observer who, over time
and in various contextually specific situations, takes on increasing degrees of responsibility
under the guidance of a master. This very model is the one which has dominated surgical and
interventional radiology skills teaching and continues to be the predominant method to date.
MOTOR SKILL DEVELOPMENT Behavioral psychologists divide motor skills learning into three distinct levels.30
Psychomotor skills are those which, after numerous repetitions, may be partially or fully
automated by the motor cortex to a level of unconsciousness.31 A common example is the
ability to ride a bicycle. Enormous amounts of energy must be used to coordinate muscle
efforts to attain smooth, steady cycling at the beginning of the learning curve. Once these
processes are internalized, one hardly needs to think about the actions. Instead, the actions are
automated freeing larger amounts of working memory for other endeavors. Basic catheter
manipulation skills in the endovascular suit fall into this skill category.
Procedural skills entail the learning of rules and/or steps. A practical example would
be a cake recipe with instructions. One cannot simply add all ingredients into a large dish,
throw it into the oven and expect a three-layered butter-cream masterpiece. Specific steps
must be followed in a particular order to achieve the desired outcome. Similar logic applies in
the human catheterization laboratory. One cannot hope to put a patient, the correct
interventional tools and a novice into a room and expect satisfactory results. The procedure
must proceed according to protocol if the patient is to benefit from treatment.
Lastly, cognitive skills encompass decision-making and feats of manual dexterity
when faced with an unfamiliar or unexpected environment. In short, these are reactions to a
new situation that are created and executed to attain a desired result based upon that
particular individual’s body of knowledge. This skill answers the question: Given an
unknown, what would one do next? In surgical specialties, a novice’s ability to correctly react
to sudden, intraoperative complications is a defining step toward mastering their trade.
13
ENDOVASCULAR SKILLS Although the exact definitions of what constitutes a skilled interventionalist remains
unwritten, a behavioral psychologist approach seems most appropriate.30,32 Psychomotor skill
acquisition, sometimes termed generation, is the process of initial learning much like learning
to ride a bicycle. During these fragile first steps, one needs to focus large amounts of
concentration and expend tremendous effort to force the body to make the new motions in an
appropriated fashion as instructed. Acquisition of the new psychomotor skill is a separate
entity from practicing that skill until mastery.19 The time from initial learning until mastery is
usually referred to as the learning curve and, unfortunately, represents the period of time
when most surgical errors are likely to happen.2,33,34
Traditional surgical skills allow the natural use of the senses to see, touch and smell as
the maneuvers are being attempted. However, endovascular intervention robs the surgeon of
the ability to see in three dimensions and to make use of the direct tactile feedback, known as
haptics. Two dimensional fluoroscopy replaces the open wound and the interventional haptic
feedback is miniscule by comparison. Thus, interventional psychomotor skills cannot be
considered interchangeable with open surgical skills or laparoscopic ones.
Procedural skills represent the ability to follow a given procedures protocol closely. In
essence, it is the ordered steps needed to perform a dance properly. Due to the nature of
interventional techniques, e.g. singular femoral artery access, the methodical introduction of
equipment through the arteriotomy according to protocol allows successful results. Failure to
do so may render hours of preparatory work meaningless as the necessary instrument may not
be able to be deployed properly causing lost Cath-Lab time, wasted equipment and extended
risk exposure for the patient. Literally speaking, there is no room for error.
Lastly, cognitive behavior denotes how a person reacts, based on their inherent body
of knowledge, when they meet with the unexpected. Simply stated, cognitive skills come into
play when surgical difficulties or complications present themselves. For example, the
textbook anatomy expected may actually be an unrecognizable congenital anomaly which
forces the interventionalists to find another vascular route to the target site impromptu.
Cognitive skills can only be fully developed in a real catheterization laboratory, human or
otherwise, because, although it may one day be possible to incorporate such scenarios into the
simulation milieu, the extent of unpredictability found in a real cath-lab includes too many
variables to be included. Thus, while procedural complications, e.g. angioplastic balloon
rupture, could be included into the simulator, human factors causing disturbances in work
14
flow such as a non-cooperative patient who refuses to lay still during angiography or
experiences nausea and vomiting during the most critical point in a procedure may not be as
readily simulated.
Skills Assessment
Who, when and what should be the final judge of a fellow’s readiness to enter the
catheterization laboratory? Three paths are available. Traditionally, subjective approval from
one’s mentor marked the passage from apprenticeship to journeyman, i.e. independence in
the Cath-Lab. While rooted in tradition, the major drawbacks of this method are its continued
use of patients to train and dependence upon a random exposure to procedures. Such a
practice is especially questionable when examined in the light of ethics-based medicine and
the increasingly litigious atmosphere in which we work.
In contrast, validated metric based virtual reality assessments offer a completely
objective, standardized model wherein fellows might attain metric benchmarks prior to
independent catheterization laboratory work.19,35,36 This seems promising, but places the
grave decision-making responsibility on microprocessors incapable of human subtleties.
Additionally, this requires reproducible metric validations for each parameter measured by
the simulator, for each module and for every type of simulator.37 Lastly, simulator metrics
assesses virtual reality skills, which are of unknown real world merit unless that particular
simulator’s metrics have been demonstrated to bestow significant benefit in the OR.10,38,39
While a simulator with construct validity can measure performance differences between
novices and experts, construct validity represents only a step towards demonstrating a
simulator’s clinical worth.
The last choice follows the path of moderation. Endovascular procedures, like all
surgical endeavors, are not merely psychomotor skills isolated from subjective judgments.
Rather, they are a marriage of the two. Thus, intervention is the application of physical skills
utilized in an intelligent manner to produce therapeutic results in unpredictable clinical
situations. Synergizing subjectivity and objectivity results in a reliable form of skill
evaluations. Similar logic led to the development of Objective Structured Assessment of
Technical Skills (OSATS), which give the evaluating proctor a validated tool to use when
assessing surgical trainees.40-44
15
Skills Transfer
Do skills learned in the virtual environment transfer to the catheterization
laboratory?45-47 The answer to this question is the gold standard which the VR-Lab must meet
in order to attain widespread scientific acceptance.13,48 Virtual reality simulators from other
disciplines have shown some predictive value, but those instances are few according to
experts.49,50 Furthermore, although assumed, no specific evidence exists for endovascular
simulators demonstrating virtual skills transferability.50,51 Early attempts to demonstrate
transferability of VR skills in other surgical disciplines have often lacked comparable training
in the control group, yet have not resulted in definitive evidence as recently pointed out by
Schijven et al.12,39,52,53
Hence, acceptance of the VR-lab as a venue for skills acquisition within the scientific
community awaits evidence demonstrating skills transfer from VR to OR.32,45-47,50 As rational
as this requirement is, the same demands were never placed upon the P-Lab. Remarkably, no
evidence supporting or refuting endovascular skills transfer from the research animal model
to the operating room (OR) or Cath-Lab exists at the time of this writing. The lack of
evidence demonstrating the transferability of skills learned in the P-Lab to the OR leads one
to conclude that the gold standard—for non-clinical endovascular interventional training—
has never been thoroughly tested in a manner befitting the scientific method.
Cost Effectiveness
The bottom line of the yearly budget has an educational impact on all institutions of
higher learning. The demands of training increasingly complex surgical procedures have risen
as the available amount of legal working hours has diminished.54-58 Given that two
alternative methods produce similar clinical results, institutions should prefer the economical
method over the more expensive one. The ethical disbursement of economical resources—
defined here as the highest pragmatic gain from the least financial burden—are a necessity in
modern healthcare. We must attempt to provide the best healthcare to the greatest number of
individuals using the finite quantities of public or private capital.
16
Ethics
Since their introduction in 1959, scientists have been striving to adhere to the three
“R’s” of replacement, reduction and refinement wherever the use of research animals are
required by experimental design.59 In their groundbreaking book, Russell and Burch urged
scientists to replace animals with insentient material, i.e. in vitro, or to substitute to a lower
species. Furthermore, they plead for a reduction in the number of animals used to obtain the
necessary information to the lowest level deemed statistically adequate. Finally, study
methodology was to be refined in such a way as to decrease the incidence and severity of
pain and distress in any animals that were to be used. Indeed, progress has been made thanks
to advances in biomedical cell biology allowing increase use of cell cultures, improved
statistical power calculation abilities and increased ethical awareness by researchers when
designing studies.
Since then, animal rights activists have been successful in lobbying for more
restrictive laws with help from a sympathizing media and an empathizing public. The
changes and restrictions brought about by their success are not necessarily deleterious to
research. The widespread use of animal care and ethical review boards are prime examples of
such progress which justly requires, at the very least, reflection about what is to be gained
scientifically from a proposed experiment and how it is to be accomplished.60 Nevertheless,
the experimental use of animals is—despite increased public awareness—frequently accepted
as scientific dogma and a necessary evil.61 Most people generally accept the pragmatic view
of mankind as the most valuable species of all. DeGrazia successfully placed both sides of the
animal rights argument on common ground when he listed a collection of points on which
both sides might agree. He stated that the use of sentient animals, i.e. those capable of
experiencing pain and distress, for medical research raises ethical issues and that these
animals deserve special care.62 Regan presented a similar argument when he stated that an
individual animal’s inherent value does not disappear merely because researchers fail to find
an alternative experimental design.63 In essence, we ought to be thinking of better ways to get
the experimental data that humanity needs while including the earlier mentioned three R’s at
each step.
Many universities provide endovascular training courses using anesthetized research
animals to help aspirants over the steepest part of the learning curve-- the window of (in-)
opportunity where most surgical errors occur. The basis for research animal training is to
maximize patient safety until the novice gains a rudimentary understanding of interventional
17
skills. While curricula vary, a normal course includes didactics followed by hands-on practice
in the research animal catheterization laboratory. The adult swine vascular tree approximates
human vessels despite anatomical incongruence and lack of pathological disease found in
patients. This method offers the advantages of in vivo training and the use of the exact
equipment with which expertise are being sought. Additionally, encountered complications
serve as an enriching experience for trainees as they learn to deal with the unexpected.
Virtual reality offers in vitro training using slightly modified equipment to treat computer
simulated patients with common human pathologies, e.g. carotid arterial stenosis.
SIMULATOR VALIDITY
A critical element of any educational measurement tool, is to assure its validity. The
most rudimentary form of validity entails face validity. Face validity is the degree of realism
the virtual simulation can mimic. Normally, subject matter experts in the specialty of interest
are allowed to subjectively evaluate how well a simulated scenario compares to their real
clinical experiences.
An obligate part of confirming a simulator’s validity is to establish from a
psychometric perspective its construct validity; defined as the ability of a tool to measure the
trait it purports to measure. Construct validity is often affirmed and inferred, by establishing
that performance improves with experience. For example, the Minimally Invasive Surgical
Trainer-Virtual Reality (MIST-VR™) construct validity was confirmed in a study which
demonstrated its ability to stratify laparoscopic VR performance based upon individual
clinical experience (n=41) using a common procedure, but nevertheless a demanding one.64
Additionally, many modern endoscopic simulators have gone through generations of
improvements and scientific evaluations to make them valuable tools for shortening the
learning curve.
A randomized, double blind study conducted by Grantcharov and coworkers using the
MIST-VR™ showed decreased surgical errors (p = 0.003) and operative times (p = 0.021)
during laparoscopic cholecystectomies for VR trained residents when compared to the control
group (n=16).39 A previous randomized, double blind study done by Seymour et al
demonstrated that VR trained surgeons completed laparoscopic cholecystectomies 29% faster
(p = 0.039) while non-VR trained surgeons were five times more likely to make errors (p =
0.039, n=16).65 Furthermore, two separate studies recorded shorter learning curves in the
acquisition of the basic psychomotor skills necessary for laparoscopic surgery when training
included VR simulators.20,66 18
A high degree of correlation between clinical skills and virtual reality skills is the
definition of concurrent validity. A strong correlation between the two measurements makes
non-clinical skills assessment and improvement possible in the VR-Lab. At present, there is
no generally accepted measurement standard of endovascular interventional skills in the
human Cath-Lab skills. Therefore, establishing true concurrent validity is impossible for the
moment because there exists nothing to compare new clinical scales against.
Clearly, the advantages and possibilities of VR training have been proven in other
surgically focused fields, but, as of yet, few studies have been focused on interventional
radiology (IR). The need exists to investigate if current VR technology is capable of
assessing the psychomotor skills of the interventionalist as has been done in laparoscopic
surgery.31 If such assessment is possible and the construct validity is verified, then progress
towards establishing trainee benchmarks can be begun.67
19
“ GOOD JUDGEMENT COMES FROM EXPERIENCE.
EXPERIENCE COMES FROM BAD JUDGEMENT.” -UNKNOWN
CATCH 22
Novice interventionalists are thus trapped in a classic Catch 22 proposition; to become
proficient in the shortest amount of time at a reasonable price for their respective educational
institutions while creating minimal ethical dissonance both in clinical and non-clinical
situations. Sound critique, either subjective or objective, and guidance from clinical mentors
speeds this journey, yet only independent time in the catheterization laboratory will lead to
the vast experience one needs to truly achieve excellence. Table 1 represents the possible
venues for this training and experience. Each has its own merits and possibilities, yet many
questions remain.
CATH-LAB P-LAB VR-LAB
Psychomotor skills training
Procedural skills training
Cognitive skills training ?
Objective feedback (OSATS or Metrics)
Subjective feedback
Experience with human pathology
Patients spared initial learning curve
Flexible learning time removed from the clinics
Reduced radiation exposure
Systematic procedural exposure
Cost effective ? ? ?
Ethical Concern ? ?
Shorter learning curves and reduced error rates ? ? ?
Maximized interventional skill training ? ? ?
= Feasible, = Infeasible, ? = Unknown
Table 1 Comparison for the alternative endovascular interventional training methods; Catheterization- (Cath-
Lab), Porcine- (P-Lab) and Virtual Reality catheterization laboratory (VR-Lab).
20
AIMS
I. To assess the construct validity of the renal artery stenosis modules of the Procedicus-
VIST™ simulator. A) Can the VR simulator stratify interventional performances based
upon prior endovascular experience? B) Will the system work as a performance
assessment tool outside the real catheterization laboratory? C) Is the VR-Lab useful as a
pedagogic tool?
II. To compare the training effects of using the porcine model to virtual reality training in
the endovascular novice. A) Is virtual reality simulation as effective a training tool as
the porcine laboratory? B) Do skills learned in virtual reality transfer to the
catheterization laboratory? C) Is one form of non-clinical training subjectively preferred
over the other by trainees?
III. To conduct an economic analysis of the two non-clinical training forms for use in basic
endovascular skills training. A) How does the VR-Lab purchase compare financially to
the renting of the P-Lab? B) How does the rental of both laboratories affect the relative
cost ratios?
IV. To assess the construct validity of the carotid artery modules of the Procedicus-VIST™.
A) Can objective metric data from these modules stratify virtual performances based
upon experience level? B) Can these simulator modules be used to assess endovascular
skills outside of the catheterization laboratory? C) Is the VR-Lab useful as an
educational tool?
21
METHODS AND MATERIAL
SUBJECTS
For study I, eight interventional radiologists and eight medical students undergoing
their surgical clerkships participated. The expert group had a mean age of 48 years (range 42-
63) and consisted of seven males and one female. Their mean IR experience was 10 years
(range 8 months - 20 years). Their mean number of renal artery stenosis (RAS) interventions
per year was 11 procedures per year (range 1-40) and two had prior simulator experience.
Four played video games “sometimes” and four played “never” on a scale that included:
often, sometimes and never. The novice group had a mean age of 28 years (range 24-32) and
also consisted of seven males and one female. No one had interventional or previous
simulator experience. Six played video games “sometimes” and two played “never.”
Study II enlisted a group of twelve vascular surgeons and interventional radiologists
to participate in a two day experimental endovascular training course consisting of the
porcine model and VR simulations. The trainees (11 males, one female, 27-61 years old) had
a mean open surgical experience of 8 years (range 0-31) and mean endovascular experience
of one year (range 0-5 years). Two participants had limited prior virtual reality simulation
exposure (<15 minutes). A group of six experienced interventional radiologists with a mean
interventional experience of 10.5 years were recruited to function as onsite proctors. Each
proctor evaluated the same trainees in the VR-Lab and the P-Lab to minimize variability.
Two highly experienced interventional radiologists with a mean of 22.5 years of
interventional experience were recruited to serve as the video assessor panel.
Figure 1 Procedicus-VIST, Mentice Medical Simulations, Gothenburg, Sweden
23
1. Insert a 0.035” J-profile guide-wire and a 4/5 F Pigtail diagnostic catheter over the 0.035”
guide-wire into the distal aorta.
2. Connect the contrast line and perform a Digital Subtraction Angiography (DSA) of the iliac
arteries via the Pigtail.
3. Use the 0.035” guidewire and Pigtail to canalize the contralateral common iliac artery. (You
may need to change catheters, e.g. Sim1 or SHK1, or guidewires, e.g. hydrophilic or stiff, in
order to accomplish this.)
4. Insert a 6/7 F guiding catheter into the contralateral common iliac artery proximally to the
stenosis.
5. Carefully transverse the “stenosis” with the guidewire.
6. With the guidewire still in position, connect the contrast line, perform a selective DSA or
roadmap of the contralateral external iliac artery.
7. Measure and evaluate the external iliac artery and the “stenosis.”
8. Insert an appropriately sized peripheral stent catheter over the wire.
9. Center the stent within the lesion and carefully deploy the stent.
10. Maintain your distal guidewire position and remove the stent catheter.
11. Insert an appropriately sized peripheral dilation balloon catheter over the 0.035” guidewire,
advance into the stent’s lumen and perform a post-deployment PTA.
12. Maintain the distal guidewire position and remove the angioplasty catheter.
13. Connect the contrast line and perform a control DSA via the guiding catheter.
14. Withdraw the 0.035” wire and introducer.
Table 2 Iliac artery stenting protocol.
Study III was conducted using economic data extracted from Study II but included no
new participants.
Study IV comprised experienced interventionalists and medical students during their
surgical clerkship. The expert group had a mean age of 49 years (range 36-65) and consisted
of fifteen males and one female. Their mean IR experience was 11 years (range 1-25). The
novice group had a mean age of 29 years (range 23-39) and consisted of thirteen males and
three females. Five had limited previous simulator experience yet none had any IR
experience. Neither group had any experience in placing carotid artery stents.
24
1. Place .035” J-type guidewire & .035” Pigtail catheter into the ascending aorta.
2. Position C-arm in LAO (25°-60°), remove .035” guidewire and perform DSA with AP and lateral
views.
3. Replace .035” wire, exchange to .035” diagnostic catheter and position the guidewire in the external
carotid artery.
4. Insert a 6F guide catheter proximally to the CCA bifurcation and remove the .035” guidewire.
5. Insert the .014” EPD guidewire, transverse the stenosis and deploy the EPD filter.
6. Using a .014” peripheral balloon catheter pre-dilate the lesion.
7. Insert .014” carotid stent catheter, deploy the stent and perform a DSA.
8. Repeat PTA if necessary.
9. Use a .014” recovery sheath catheter to collect the EPD.
10. Perform control DSA of the right ICA including intracranial views.
11. Check for spasm? Dissection? Remove all equipment.
Table 3 Carotid artery stenting protocol for the right internal carotid artery using a femoral artery approach.
VIRTUAL REALITY LABORATORY
The virtual reality simulator used in all experiments was the Procedicus-VIST™
system (Mentice Medical Simulations, Gothenburg, Sweden) which consisted of a double
processor computer, a touch-sensitive screen, a viewing screen and a simulator dummy as
seen in Figure 1. The dummy concealed various mechanical systems, which registered the
physical movements of and produced the haptic feedback to the users. The touch screen was
driven by a menu system which allowed the trainees to select guide wire, diagnostic catheter,
guiding catheter or stent/balloon catheter by type and diameter. Fluoroscopic view,
instrument selection, total IV contrast dose and intervention time were continuously
displayed on the viewing screen. The simulator dummy consisted of a plastic human form in
the supine position with a right femoral artery port lying upon a catheterization laboratory
type table. An introducer was permanently placed within the right femoral artery. A standard
two-pedal system with X-Ray and Cine lay under the table. A double joystick control box
allowed the operator to; 1) Position the virtual fluoroscope 2) Zoom the view 3) Replay cine
sequences 4) Capture and simultaneously display an overlapping roadmap on the viewing 25
screen 5) Reposition the catheterization table virtually. An inflator apparatus with a pressure
gauge was permanently attached to the main system. When either a balloon or stent catheter
was selected and introduced into the dummy, the inflator produced appropriate morphological
changes on screen. Once deployed, stents remained in place for the entire sequence. A
permanently attached IV contrast syringe prepared with a one-way valve was used for
simulated contrast infusion. Replenishment was accomplished simply by retracting the
plunger, which drew air in place of real contrast into the syringe. Real catheters and
guidewires in sizes ranging from 0.014”-0.035” (5-7 French (Fr)) were used. The
profile/flexible ends could not be inserted into the machine due to the simulator’s mechanical
design. Instead, straight ends were used in order to properly engage the haptic mechanisms
within the device. The program automatically displayed the correct tip profile for the selected
instrument on the viewing screen.
PORCINE LABORATORY
In three separate catheterization laboratories, normally fed Swedish swine were
sedated with i.m. ketamine and dormicum and ventilated on a mixture of oxygen and N2O.
Anesthesia was maintained with continuous i.v. infusion of thiopenthatol and buprenorphine
supplemented with isoflurane as necessary. After catheterization of the right femoral artery
using the Seldinger technique and a 10Fr introducer sheath (Cook Inc, USA), an i.v. bolus
dose of 300U/kg of Heparin was given. Repeat anticoagulation was given every 60 minutes
thereafter. Preparation for iliac artery stenting was accomplished using various equipment
from Cordis, Johnson & Johnson, USA and Boston Scientific, USA. The final stenting and
angioplasty was performed with a nitinol stent and an optiplast balloon from Bard, Inc, USA
(Luminex® and XT Optiplast®). Fluoroscopy and Digital Subtraction Angiography (DSA)
was performed using Philips and GE fluoroscopic equipment (Philips Medical Systems,
Eindhoven, The Netherlands and General Electric Healthcare, UK). Using sodium ioxaglat
(Omnipaque® 240 ml) as the contrast medium, the size of the left external iliac artery and the
“stenosis” to be treated was estimated using a semi-quantitative method where the guiding
catheter was used as a reference. The “stenosis” was delineated by the proctor on the
fluoroscopic screen using transparent, self-adhesive plastic book marking tabs. At the
conclusion of the experiment the pigs were euthanized with a lethal IV bolus of potassium
chloride.
26
LOGISTICS
Studies I & IV
Prior to construct validations, all participants received a 45 minute, standardized
didactic introduction to the simulator and the requisite endovascular techniques. Participants
were given the same modules to complete, the rules for completion and the role of the
proctors. The objective of the procedure was to revascularize the afflicted artery using either
balloon angioplasty, stents or a combination of both. Study I used six different renal artery
stenosis (RAS) modules, i.e. simulated patients, which were completed twice without the aid
of embolic protection devices, once during familiarization and again during the testing phase.
Study IV assessed performance using a unique carotid artery stenosis (CAS) module, i.e. it
was not available during familiarization training, with the aid of an embolic protection
device. Any case which took over 30 minutes (I) or 60 minutes (IV) to complete was
abandoned, during familiarization and testing, to allow attempts at the remaining cases. A
short journal was presented onscreen and the intervention timer automatically started once the
operator began instrument selection. The operators were requested not to be overly concerned
with speed and to complete each procedure to the best of their ability. All subjects attempted
each procedure twice, once during the familiarization period and again during an undisturbed
test period. In order to mitigate knowledge-based bias, full disclosure of the performance
metrics to be recorded was given to both groups, as it was assumed that experts, but not
novices, might have had prior knowledge of important metrics within interventional
radiology procedures.
Study II
Trainees were informed that they would be participating in an experiment involving
proctored evaluations and randomized training methods. Prior to arrival, each received an
information package containing the iliac artery stenosis (IAS) procedure protocol,
instructions regarding their responsibilities during the experiment and a demographic
questionnaire used to experience-stratify and randomize the trainees into four alternative
training groups. Proctors and video assessors received a one hour introduction to the trainee
evaluation methods prior to the study’s commencement. All participants received a one hour
didactic introduction to the simulator and the necessary endovascular techniques before the
evaluations began. The interventional objective was to revascularize the iliac artery with
27
balloon angioplasty and nitinol stents using the standardized protocol, Table 2. Any case
which took over 30 minutes to complete was aborted. Following initial evaluations, training
groups completed two unevaluated training sessions of three hours each according to their
randomized method(s), Figure 2.
28
Stratification & Didactics
Proctor evaluations P Lab & VR
Figure 2 Study II, training flowchart.
Study III
Actual cost data were collected from the university P-Lab and VR-Lab during the two
day course conducted in October 2005. Rental fees for both labs included full ancillary
support for the swine and simulators. Proctors and video assessor salaries were not included
in this analysis as teaching was assumed to be a regular function of an academic institution.
Each proctor attended both days of the course, totaling 16 hours per course, while the video
assessors took one eight hour day to complete video recording reviews. Market research and
formal manufacturer quotes were used to complete any missing prices. The costs of the VR-
Lab was chosen as the numerator and the P-Lab as the denominator for the comparison ratio.
A five year analysis was chosen as the longitudinal length of comparison to reflect the
expected life cycle of the simulator.68 The original value of the Procedicus-VIST™ was
assumed to be €200,000. Residual value was presumed to 25% of the original value, i.e.
P - Lab
P - Lab
P - Lab
VR - Lab P - Lab VR-Lab
VR-Lab VR - Lab
- -Lab
Proctor evaluations P-Lab & VR-Lab
€50,000, after the five year life cycle ended making the depreciation value equal to €150,000
with an annual depreciation value of €30,000 using straight-line depreciation. Laboratory
rental prices were assumed to increase at a rate of 2.5% per year to match the current Euro
inflation rate. Annual national IR and vascular surgical training demand was calculated as 52
and estimated to increase at a 2% per year. Sensitivity analysis— an essential tool used in
making economical decisions to clarify the impact of price variability—was performed to
assess the impact of price variations for each laboratory given a 50% rise or fall in costs.
Such an analysis helps safeguard against making incorrect decisions based upon singularly
interpreted financial data. The first year annual difference was assumed as the short-term
potential savings and the five year cumulative total as the long-term potential savings.
ENDPOINTS
Objective Endpoints
Objective performance metrics for studies I and IV were automatically recorded for
each case during the testing phase only by the Procedicus-VIST™ software, Table 4. The
Total Score performance evaluation forms used in study II consisted of modified surgical
evaluation forms originally developed for surgical skills assessment.38. The iliac artery
procedure forms used were the Task Specific Checklist (TSC) and the Global Rating Scale
(GRS), which yielded a Total score. The TSC was completed during the procedure, Table 5.
Correct completion of a task resulted in the assignment of one point (max=14). If the task
was improperly performed or omitted, the trainee received zero points. Additionally, if the
trainee asked for help or the proctor needed to intervene to prevent a catastrophic event, the
trainee received a zero for that task, even if it was completed properly afterwards. The GRS,
based on multiple questions and a five point anchored Lichert scale, was filled out once the
procedure was completed to ensure an overview of the entire performance sequence
(max=45), Table 6. Proctors performed the evaluations during and immediately after the
procedures were completed. Video assessors evaluations, performed using blinded video
recordings from the VR-Lab, were used to calculate Total Score, Task Specific Checklist and
Global Rating Scale inter-rater reliability. Blinded video ensured that each participant’s
identity was protected during videography.
29
Procedure Time Minutes needed to complete the entire procedure
Fluoroscope Time Minutes the fluoroscope was used during the procedure
Contrast Total amount of contrast medium used, in milliliters
Cine Loops Number recorded during procedure
Lesion Coverage % Percentage of lesion covered by selected tool
Tool : Lesion Ratio Inflated tool’s diameter : lesion’s diameter
Placement Accuracy Distance, in millimeters, from the actual placement of the tool
to the lesion’s center, longitudinally
Residual Stenosis Percentage stenosis post PTA or stent deployment
Table 4 Metric definitions.
OMITTED OR
INCORRECT
DONE
CORRECTLY
INSTRUCTION
REQUIRED*
1. Positioned guidewire and diagnostic catheter
correctly? 0 1
2. Distal aorta DSA conducted correctly? 0 1
3. Proper guidewire technique used to gain contralateral
access? 0 1
4. Guiding catheter properly placed in the contralateral
common iliac artery? 0 1
5. Transversed “stenosis” with correct technique? 0 1
6. External iliac artery DSA / roadmap conducted
properly? 0 1
7. Measurements and evaluations performed accurately? 0 1
8. Proper stent catheter selection? 0 1
9. Deployed stent accurately? 0 1
10. Maintained guidewire position across lesion? 0 1
11. PTA conducted correctly? 0 1
12. Maintained guidewire position across lesion? 0 1
13. Control DSA conducted correctly? 0 1
14. Extraction of guidewire and guiding catheter
performed correctly? 0 1
Table 5 Iliac artery stenosis task specific checklist. * Receiving instruction resulted in a zero even if the step
was correctly executed afterward.
30
ASEPTIC TECHNIQUE* 1 2 3 4 5
Sloppy with high risk for
contamination
Reasonable but some lapses that risk loss
of sterility
Careful with little risk of
compromising sterility
RESPECT FOR TISSUE 1 2 3 4 5
Frequently uses unnecessary force or
causes damage
Careful handling but occasionally causes inadvertent damage
Consistently handles
with minimal damage
FLUOROSCOPIC PROFICIENCY 1 2 3 4 5
Poor control and selection of
inappropriate view or causes patient
injury
Competent use but with some lapses in
control or sub-optimal views
Prompt attainment of
appropriate fluoroscopic views
TIME & MOTION 1 2 3 4 5
Slow with many unnecessary moves
and instrument changes
Makes reasonable progress but some unnecessary moves
Clear economy of
movement and maximum efficiency
INSTRUMENT HANDLING & SAFETY 1 2 3 4 5
Repeatedly makes tentative, awkward
or unsafe moves
Competent use but occasionally
awkward or tentative Fluid movements
without stiffness
KNOWLEDGE OF INSTRUMENTS* 1 2 3 4 5
Frequently asks for or uses wrong
instrument
Knows names of most instruments and uses them properly
Obviously familiar with all instruments
and their uses KNOWLEDGE OF SPECIFIC PROCEDURE*
1 2 3 4 5 Require specific
instruction for most steps
Knows all the important steps
Demonstrates familiarity with all
steps QUALITY OF FINAL PRODUCT
1 2 3 4 5
Well below standard and likely to fail
Deficiencies but would probably
function adequately
Excellent with no flaws and likely to
function well RADIATION DISCIPLINE
1 2 3 4 5 Patient and Operator
repeatedly
overexposed.
Poor use of shutters
and shielding.
Aware of exposure
but needs
improvement.
Adequate shutter and
shield use
Judicious use of
fluoroscopy.
Exposure kept to a
bare minimum.
Table 6 Iliac artery stenosis global rating scale. * Not possible to evaluate on video.
31
Subjective Endpoints
Participants completed exit surveys to record their demographics and subjective
opinions regarding the simulator and training formats. For study I, the questionnaire required
the participants to use a visual analogue scale (VAS) to rate the following eight parameters:
1) Total realism 2) Guidewire realism 3) Catheter realism 4) Balloon realism 5) Stent realism
6) Fluoroscopic realism 7) Joystick control realism and 8) Pedagogic instrument. VAS scores
marked on a 10 centimeter line ranged from zero to 100, where a higher score meant a better
rating. For studies II and IV, trainees completed exit surveys, again using the VAS system,
ranging from zero (strongly disagree) to 100 (strongly agree) to indicate their agreement to
statements regarding the two training methods, Table 7.
VR training is better than the porcine lab
I understand the IAS procedure better than I did before the course
VR training could replace the porcine lab
VR simulation training should be obligatory
Porcine lab training should be obligatory
I could have completed the procedure equally well without the Proctor
VR is a valuable training instrument
The VR simulations were realistic compared to the porcine lab
Porcine lab training is better than the VR lab
Table 7 Exit survey statements.
STATISTICAL ANALYSIS
Data analysis was performed using the statistical program “R”, an open-sourced
language statistical computing and graphics program.69 For studies I and IV, the difference of
interest was defined as 2 SD from the expert mean for each objective parameter. The large
difference was used to detect a presumably vast performance difference between the groups
based upon their endovascular experience levels and to maintain a small study size. Only
results from the testing period were analyzed. No analysis for correlation for video gaming
was performed. Student’s t-test was used for analysis of the metric means. Study II included
an analysis of covariance (ANCOVA) to assess the effect of the independent variables--
previous number of P-Lab/VR-Lab sessions, interventional experience and surgical
experience-- on the dependent variable Total-score. The dependence of the multiple
32
observations for each person, i.e. clustered data, was accounted for through the use of
generalized estimating equations (GEE).70 GEE provided adjusted standard errors for
evaluating significance of terms in the ANCOVA models. Video assessor evaluations from
the VR-Lab were compared to the proctor evaluations to establish inter rater reliability
(IRR).42,71 Statistical significance was assumed to be present when p < 0.05. Study III ‘s
economical analysis was conducted according to the National Institute of Health’s
recommendations for the implementation of information technology (IT) products.68
Economical sensitivity analysis consisted of increasing or decreasing the estimated costs for
each laboratory by 50% to produce varying cost ratios.
ETHICAL APPROVAL
As studies I, III and IV were conducted outside of the research and human
catheterization laboratories, no ethical review board decision was sought. Study II was
approved by the local ethics review board and conformed to the Guide for the Care and Use
of Laboratory Animals published by the U.S. National Institute of Health (NIH).72
33
RESULTS
PAPER I
Table 8 contains the objective metric performance data for each parameter by group
for each of the six cases. Statistical analysis across the different modules revealed no
significant differences in performances between the two groups in procedure time, number of
cine loops, lesion coverage, tool to lesion ratio, placement accuracy and residual stenosis.
The total fluoroscopic use was greater for the novice group (p < 0.01). One expert failed to
finish a module within the 30 minute cutoff limit. All others participants completed the
assigned modules on time.
CASE 1 CASE 2 CASE 3 CASE 4 CASE 5 CASE 6
E 8.4 (2.1) 9.6 (4.0) 13.1 (4.1) 11.4 (2.6) 7.9 (1.8) 10.1 (2.2)
N 12.7 (4.6) 16.3 (3.7) 14.9 (5.9) 13.9 (4.0) 11.0 (3.8) 11.9 (4.0)
Procedure
Time (mins)
P <0.02 <0.01 NS NS <0.05 NS
E 3.5 (1.3) 4.4 (1.5) 5.8 (1.4) 4.9 (2.0) 3.0 (0.7) 4.7 (1.3)
N 7.7 (2.9) 9.5 (3.1) 9.1 (4.8) 9.8 (4.2) 6.4 (2.1) 7.5 (2.7)
Fluoro time
(mins)
P <0.001 <0.001 <0.05 <0.01 <0.01 <0.02
E 3.1 (1.1) 3.6 (1.6) 6.3 (3.2) 3.6 (1.3) 3.3 (1.7) 4.9 (2.4)
N 3.6 (2.6) 4.8 (3.9) 3.9 (2.0) 3.0 (2.2) 3.8 (2.8) 2.9 (1.6)
Cine loops
(n) P NS NS NS NS NS NS
E 99.2 (1.9) 94.8 (5.8) 98.0 (5.2) 83.1 (12.8) 90.9 (12.5) 58.8 (32.9)
N 97.4 (3.8) 93.4 (8.7) 94.1 (7.4) 85.6 (9.5) 95.8 (6.9) 68.5 (22.9)
Lesion
coverage (%)
P NS NS NS NS NS NS
E 100.0 (9.0) 90.0 (18.2) 109.2 (16.3) 99.6 (16.3) 87.9 (8.2) 98.3 (13.2)
N 90.3 (24.1) 87.1 (26.9) 100.3 (23.4) 91.4 (16.3) 84.4 (11.7) 84.1 (14.7)
Tool : vessel
ratio (%)
P NS NS NS NS NS NS
E 1.0 (1.2) 2.3 (1.8) 1.2 (1.2) 3.0 (1.6) 1.9 (1.2) 6.8 (5.3)
N 1.5 (0.7) 2.1 (1.7) 1.8 (1.1) 2.9 (0.8) 1.3 (1.3) 4.6 (3.6)
Placement
accuracy
(mm) P NS NS NS NS NS NS
E 2.8 (4.3) 14.1 (12.8) 4.7 (6.4) 6.7 (8.9) 12.9 (6.2) 5.8 (8.6)
N 13.6 (20.7) 17.7 (21.5) 11.1 (14.5) 11.9 (12.9) 15.7 (11.4) 16.1 (14.5)
Residual
stenosis (%)
P NS NS NS NS NS NS
Table 8 Study I, means and standard deviations (SD) for experts (E) and novices (N) with p-values (P). NS =
not significant.
35
Table 9 contains the subjective rating data from the exit survey. The mean VAS
scores rating the use of the VIST™ as a pedagogic instrument was 89 (SD 9) for the experts
and 92 (SD 9) for the novices. Statistical analysis revealed no significant differences in mean
VAS scores for total realism, guidewire realism, catheter realism, balloon realism, stent
realism, fluoroscopic realism and use as a pedagogic instrument. Mean VAS scores for
joystick realism was rated significantly higher by the expert group (p < 0.02).
EXPERTS NOVICES P-VALUE
Total realism 61 (27) 68 (16) NS
Guidewire realism 71 (29) 85 (9) NS
Catheter realism 70 (30) 83 (13) NS
Balloon realism 75 (24) 82 (12) NS
Stent realism 59 (34) 77 (12) NS
Fluoroscopic realism 95 (9) 86 (10) NS
Joystick realism 94 (7) 71 (24) < 0.02
Pedagogic instrument 91 (9) 92 (9) NS
Table 9 Study I, VAS scores (0-100) with means and standard deviations (SD).
PAPER II
36
Total Score IRR, determined by calculating the inter-rater correlation coefficient
(ICC) between the proctors and video assessors, was found to be substantial at a value of
0.679. Cumulative VR-Lab sessions improved VR-Lab Total scores (β = 3.0, p = 0.0015) and
P-Lab Total scores (β = 1.8, p = 0.0452), Table 10. P-Lab sessions improved P-Lab Total
scores (β = 4.1, p < 0.0001) but had no effect on VR-Lab Total scores. In the general
statistical model for Total scores from all laboratories, both P-Lab sessions (β = 2.6, p =
0.0010) and VR-Lab sessions (β = 2.4, p = 0.0032) significantly improved Total scores.
Neither previous surgical nor IR experience affected Total scores. VR-Lab Total scores were
consistently higher than P-Lab scores (∆ = 6.7, p < 0.0001). Analysis of mean Total scores in
each laboratory failed to detect any ceiling effect, Figure 3.
TOTAL SCORE Β 95% CI P-VALUE
Surgical Experience - 0.2 - 0.5, 0.1 0.1141
IR Experience 1.4 - 0.8, 3.6 0.2295
P-Lab sessions 2.6 1.0, 4.1 0.0010
VR-Lab sessions 2.4 0.8, 4.1 0.0032
VR-LAB TOTAL SCORES
Surgical Experience - 0.1 - 0.4, 0.1 0.2612
IR Experience 1.6 - 0.3, 3.5 0.0921
P-Lab sessions 1.3 - 0.2, 2.8 0.0902
VR-Lab sessions 3.0 1.1, 4.9 0.0015
P-LAB TOTAL SCORE
Surgical Experience - 0.3 - 0.7, 0.0 0.0826
IR Experience 1.2 -1.5, 3.9 0.3923
P-Lab sessions 4.1 2.1, 6.1 < 0.0001
VR-Lab sessions 1.8 0.0, 3.9 0.0452
Table 10 Study II, Total score ANCOVA partial coefficients.
Figure 3 Study II, Total score box-whisker plots.
37
Ten aborts-- seven failures to finish (58%) and 3 dissections (25%)-- occurred during
the initial evaluations in the P-Lab (completion rate=17%). Five aborts, all failures to finish
(42%), occurred during the VR-Lab initial evaluations (completion rate=47%). Two aborts--
one perforation (8%) and one failure to finish (8%)-- occurred during the P-Lab final
evaluations (completion rate=83%). All trainees completed the IAS procedure within the
allotted time in the final evaluations in the VR-Lab.
Subjects agreed more with the statement that P-Lab training was better compared to
the VR-Lab training (mean 87, SD 17) than to the reverse statement (mean 17, SD 18), Table
11. The majority agreed that both forms of training should be obligatory; P-Lab (mean 79,
SD 31), VR-Lab (mean 79, SD 17). Realism ratings for the VR-Lab compared to the P-Lab
were low (mean 37, SD 23) and most disagreed with the statement that VR-Lab training was
capable of replacing the P-Lab (mean 15, SD 17). However, most agreed that the VR was a
valuable training instrument (mean 80, SD 25). Few agreed with the statement that they
would have been able to complete the IAS procedure equally well without a proctor (mean
17, SD 29). Finally, most agreed that they understood the procedure better after the course
than they did upon arrival (mean 81, SD 30).
MEAN (SD)
VR training is better than the porcine lab 17 (18)
I understand the IAS procedure better than I did before the course 81 (30)
VR training could replace the porcine lab 15 (17)
VR simulation training should be obligatory 79 (17)
Porcine lab training should be obligatory 79 (31)
I could have completed the procedure equally well without the Proctor 17 (29)
VR is a valuable training instrument 80 (25)
The VR simulations were realistic compared to the porcine lab 37 (23)
Porcine lab training is better than the VR lab 87 (17)
Table 11 Study II, exit survey results . VAS scores (0=strongly disagree, 100=strongly agree).
38
PAPER III
Using a VR-Lab purchase vs. P-Lab rental analysis, the cost of the VR-Lab for a two
day course given to 12 study participants was €34,348 or €2,862 per trainee, Table 12. In
contrast, the pig lab cost €46,350 to for these same 12 participants, Table 13. This translates
to an approximate cost of €3,862 per trainee. The cost ratio of the VR-Lab to the P-Lab was
found to be 0.74 in favor of the VR-Lab, Table 14. Sensitivity analysis resulted in a cost ratio
range of 0.25 in favor of the VR-Lab to 2.22 in favor of the P-Lab. The first year potential
national savings amounted to €52,009 assuming exclusive use of the VR-Lab for 52 course
participants, Table 15. At the end of the five-year, inflation-adjusted period, the cumulative
training savings totaled €325,314 excluding a residual value of €50,000 for the purchased VR
system.
Using a rental analysis, the cost of the VR-Lab for a two day course was €10,768 or
€897 per trainee. The P-Lab costs remained the same. The cost ratio of the VR-Lab to the P-
Lab was found to be 0.23 in favor of the VR-Lab. Sensitivity analysis resulted in a cost ratio
range of 0.08 to 0.70 in favor of the VR-Lab. The first year potential national savings
amounted to €154,189. At the end of the five-year, inflation-adjusted period, the cumulative
training savings totaled €844,365.
The largest item cost in the P-lab was the intravascular stents (€38.640, 83% of total)
while these constituted a relatively small amount of the VR-Lab budget (€2,140, 6% of total
if purchased or 20% using rental analysis). The largest item cost in the VR-Lab was either the
annual purchase price (€30,000, 87%) or the rental fees (€6,420, 60%) for the VIST simulator
compared to the rental fees for the P-Lab (€4,140, 9%).
39
Reusable Number Price (€) Total (€).035” Guidewire 4 37 148.014” Guidewire 4 50 200
5 F Diagnostic, short 4 33 1325 F Diagnostic, long 4 33 132
10 F Introducer 4 36 1448 F Crossover, 43 mm 4 68 2728 F Crossover, 33 mm 4 68 272
8 F Guide, 70 mm 4 50 200Luminex Stent 4 535 2,140
.035” XT PTA balloon 4 77 308
.014” XT PTA balloon 4 100 400 Subtotal 4,348
Purchase Number Annual Price (€) Total (€)Procedicus-VIST™ 1 30,000 30,000
Grand Total 34,348 Individual Cost 2,862
Rental Number Annual Price (€) Total (€)Procedicus-VIST™ 3 2,140 6,420
Grand Total 10,768 Individual Cost 897
Table 12 Study III, VR-Lab training cost per individual per course. Purchase and rental individual costs listed separately.
40
Reusable Number Price (€) Total (€)Puncture Needles 6 6 36
10F Introducer 6 36 216.035” Guidewire, 160 cm, straight 12 37 444.035” Guidewire, 260 cm, straight 6 18 108.035” Guidewire, 260 cm, curved 6 19 114
8F Balkin sheath, “Up&Over” 12 68 816.035” Terumo, 160 cm, straight 6 27 162
.035” Terumo, 160, curved 6 27 162.035” Terumo, 260 cm 6 68 408
5F Contra/”Pigtail” 12 33 3965F Cobra catheter 12 20 2404F Cobra catheter 6 20 1205F Sim 1 catheter 6 20 120
5F Bernstein catheter 6 20 120Pressure Guide Handle 6 18 108
Subtotal 1 3,570
Consumables, Evaluations Number Price (€) Total (€)Luminex, Nitinol Stent 24 535 12,840
Omnipaque contrast (100 ml) 6 83 498 Subtotal 2 13,338
Consumables, Training Number Price (€) Total (€)Luminex, Nitinol Stent 36 535 19,260
Saxx Stent 12 385 4,620Optima XT PTA balloon 12 77 924
Omnipaque contrast (100 ml) 6 83 498 Subtotal 3 25,302
Rental Fees Number Price (€) Total (€)Catheterization Laboratories 3 440 1,320
Experimental swine 6 413 2,478Infusion pumps 6 13 78
Tandem Fluoroscopy 6 44 264 Subtotal 4 4,140 Grand Total 46,350 Individual Cost 3,863
Table 13 Study III, P-Lab training cost per individual per course.
41
VR-Lab P-Lab VR-Lab Purchase
P-Lab Rental 150% VR-Lab 100% VR-Lab 50% VR-Lab
150% P-Lab 0.74 0.49 0.25
100% P-Lab 1.11 0.74 0.37
50% P-Lab 2.22 1.48 0.74
VR-Lab P-Lab VR-Lab Rental
P-Lab Rental 150% VR-Lab 100% VR-Lab 50% VR-Lab
150% P-Lab 0.23 0.15 0.08
100% P-Lab 0.35 0.23 0.12
50% P-Lab 0.70 0.46 0.23
Table 14 Sensitivity analysis. Cost ratios at various percentages of the cost estimates.
42
Lifecycle Course Demand VR-Lab Purchase P-Lab Rental Year 1 Year 2 Year 3 Year 4 Year 5
Course Participants 52 53 54 55 56
Individual Cost : VR-Lab 2,862 2,862 2,862 2,862 2,862 P-Lab 3,863 3,959 4,058 4,159 4,263
Annual Cost :
VR-Lab Total 148,841 151,818 154,855 157,952 161,111 P-Lab Total 200,850 209,989 219,543 229,532 239,976
Annual Difference 52,009 58,171 64,689 71,581 78,865 5 Year Difference 325,314
Course Demand VR-Lab Rental P-Lab Rental Year 1 Year 2 Year 3 Year 4 Year 5
Individual Cost : VR-Lab 897 920 943 966 990 P-Lab 3,863 3,959 4,058 4,159 4,263
Annual Cost :
P-Lab Total 46,661 48,784 51,004 53,325 55,751 VR-Lab Total 200,850 209,989 219,543 229,532 239,976
Annual Difference 154,189 161,204 168,539 176,208 184,225 5 Year Difference 844,365
Table 15 Five year projected national savings in Euros (€) for interventional radiologists and vascular surgeon
course using purchased or rented VR-Lab in place of P-Lab.
PAPER IV
Of the seven recorded metrics, only chronologic measurements were found to be
significantly better amongst the experts, Table 16. Procedure and fluoroscopic time was 8.7
and 8.7 minutes greater in the novice group, p = 0.0066 and p = 0.0031 respectively. There
were no significant differences in performance between the two groups for the metrics of cine
loops, tool:vessel ratio, coverage percentage, placement accuracy or residual stenosis. The
differences in the means of these performance metrics on occasion showed a tendency
43
towards a statistically significant difference. For example, experts recorded fewer cine loops,
had better residual stenosis and demonstrated better tool:vessel ratios, yet they had worse
placement accuracy and poorer lesion coverage. All participants completed the procedure in
the allotted time frame.
Contrast medium measurement metrics were found to be too imprecise for
comparative statistical analysis. Specifically, volume measurements varied widely for
variable injection rates and volumes which resulted in substantially lower cumulative
volumes than expected, Table 17.
Table 18 shows the results of the questionnaire administered to participants soliciting
their views about VIST™. Students and experts did not differ on four of ten dimensions,
including; frequency of video gaming habits, belief that the VIST™ is a good pedagogic
instrument, belief that obligatory VR-Lab training is desirable and the desire for proctored
training. Amongst the significantly different VAS score categories, novices tended to view
VR-Lab training more positively compared to reading alone. They also rated the simulations
more realistic, favored them over didactic/AV instruction and left the course with a better
understanding of the CAS procedure. However, novices would have preferred to have more
time with the simulator while the experts felt adequately familiarized. Novices expressed an
increased interest in endovascular medicine whereas the expert group did not.
EXPERTS STUDENTS P-VALUE
Total Time (min) 38.6 47.3 0.0066
Fluoro Time (min) 18.0 26.7 0.0031
Cine Loops (n) 11.0 12.4 0.3214
Placement Accuracy (mm) 4.1 2.4 0.1334
Residual Stenosis (%) 22.3 28.7 0.1405
Tool to vessel ratio 77.8 71.3 0.1356
Coverage (%) 90.8 94.8 0.5206
Table 16 Study IV, metric means compared using Student’s t-test.
44
Repeated Injection of 20 ml @ Repeated Injection of 10 ml @
20 ml/sec 5 ml/sec 10 ml/sec 5 ml/sec
Observed
Metric
Expected
Metric
Observed
Metric
Expected
Metric
Observed
Metric
Expected
Metric
Observed
Metric
Expected
Metric
4.2 20 12.6 20 2.7 10 6.4 10
8.6 40 25.6 40 5.4 20 12.9 20
12.7 60 37.4 60 8.2 30 19.3 30
17.0 80 49.9 80 10.9 40 25.8 40
21.0 100 62.5 100 13.7 50 32.1 50
25.2 120 75.0 120 16.4 60 38.5 60
29.3 140 87.5 140 19.0 70 45.0 70
33.3 160 100.0 160 21.7 80 51.5 80
37.4 180 112.6 180 24.5 90 58.0 90
Total Total Total Total Total Total Total Total
41.5 200 125.3 200 27.1 100 64.5 100
Table 17 Study IV, observed versus expected contrast medium volume measurements (ml) at varying boluses
and injection rates.
EXPERTS STUDENTS P-VALUE
I play video games often 8.8 21.0 0.0756
VR training is better than reading 70.5 93.6 0.0004
I now understand CAS better 62.4 89.7 0.0024
VR could replace didactics/AV instruction 52.3 80.1 0.0014
VR training should be obligatory 73.4 83.4 0.1521
I had adequate time with VIST 70.1 44.1 0.0074
I had no need for the facilitator 16.6 30.8 0.1191
The VIST is a good pedagogic instrument 80.4 86.4 0.2848
The simulation was realistic 67.5 82.5 0.0104
I am now more interested in endovascular medicine 38.6 80.3 0.0001
Table 18 Study IV, exit survey data. VAS scores (0 strongly disagree-100 strongly agree).
45
DISCUSSION
SKILLS ASSESSMENT
The psychometric validation of any virtual reality machine-generated metrics is
critical in order to understand the strengths and limitations of emerging educational tools.
Studies I and IV attempted to show construct validity and specifically chose two populations,
in which we anticipated large differences and hence hypothesized the need for a relatively
small sample size. This notwithstanding, essentially no differences were seen in the majority
of parameters measured between the two groups excepting fluoroscopic and procedure times.
Whether or not the experienced group’s performance failed to be superior to the
novice's due to a true lack of difference in ability or to the simulator’s inability to detect a
difference is unknown. However, we presume that—due to the tremendous difference in
endovascular experience and relative homogeneity in video gaming habits (Expert 50% vs.
Novices 75% who played “sometimes” for Study I)—there was, in fact, a difference which
was not detected. Additionally, no statistically significant difference in video gaming habits
was demonstrated in Study IV in contrast to what has been found in laparoscopic VR
simulator studies.73 Given the enormous variation found in video gaming environments and
individual game objectives, it would appear to be difficult to predict a significant impact on
VR-Lab skills independent of the particular video games the participants regularly enjoyed.74
That is, video game human-computer interactions vary to such an extent that some games
might well contribute to VR-Lab performance (ex. flight simulators) whereas others would
undoubtedly not (ex. Sims II©). Thus, the closer the visuo-spatial environment in VR
resembles the real environment, the better the transferability.47,75,76
Alternatively, the modules in question may not have been technically challenging
enough to register real skill differences. Furthermore, we noted that novices appeared very
focused on giving their best performance while the experts seemed more experimental with
the capabilities and limitations of the simulators. These differing attitudes, although not
systematically assessed, could have contributed to the equivocal objective data. Other studies
have shown very rapid learning curves for novices when their VR-Lab performances are
compared to experts and the same may have been true here.77,78 Interestingly, subjective
opinions from both groups regarding the simulator’s usefulness were positive, which is not an
unimportant fact when assessing a new training device.79 However, these two studies focused
47
on the validation of the simulator’s metrics rather than the transferability of subjective
opinions to objective performances. Again, the focus in this dissertation was on the metrics,
not the educational utility of the machine, which may very well have clinical relevance.
From the observations, it seems that basic psychomotor and procedural skills might be
possible to learn using endovascular simulations. The combination of using actual, albeit
modified, clinical equipment in conjunction with the Procedicus-VIST™ system’s haptic
feedback should, reasonably, be directly transferable to the clinical setting, i.e. the human
catheterization laboratory. Procedural skill, i.e. the actual order in which the procedure in
question should be performed, also seems feasible to learn outside of the catheterization
laboratory. Having learned these two basic skills ex vivo, i.e. in an experimental environment,
would allow the interventionalist to focus on the patient and the job at hand in vivo, i.e. in
real life. While performing in the actual catheterization laboratory will always be more
difficult than in the virtual one, the better one’s training has been, the easier it becomes to
meet real world challenges. Virtual reality training may serve as a stepping stone in attaining
procedural proficiency before one proceeds to one-on-one instruction under an experienced
interventionalist.
The deployment of virtual reality simulations and other forms of laboratory-based
training is an adjunct, not a replacement for the traditional apprenticeship model of technical
skill education. The current model suffers from a curriculum that is based on patient
availability and the logging of arbitrary numbers of cases, not the achievement of specific
objectives. Given the dramatic changes that are taking place world-wide with regards to work
hour restrictions, systems of training that afford residents the opportunities for deliberate
practice are desperately needed. The alternative might be lengthening an already too long
training process, or worse, the graduation of tomorrow’s specialist who are less well-trained
than those of today.
There is an old adage in procedural medicine that “first you get good and then you get
fast”. In that vein, chronological measurements such as procedure and fluoroscopic times
may be rudimentary measures of efficiency but, taken alone, are poor proxies for quality.
Furthermore, exemplary results often demand longer procedure times and increased
fluoroscopy to ensure exact angiographic interpretation and optimal stent placement. In
contrast, a rapid yet sloppy procedure may result in iatrogenic vessel damage,
cerebrovascular insult or persistent stenosis requiring open surgical repair.
48
General uncertainty due to such things as confusing equipment nomenclature, lack of
anatomical familiarity, awkward use of the fluoroscope and complex procedure sequence
may have caused the observed differences in speed between the experts and novices, rather
than any appreciable differences in the final quality of the interventions. If true, a longer
introduction period with the simulator and CAS training may have mitigated the observed
time differences as has been found in other studies.78,80 Interestingly, the mean differences in
fluoroscopic time equates almost exactly to the greater procedure times registered in the
novice group, 8.7 (p = 0.0066) and 8.7 (p = 0.0031) minutes. Therefore, it seems plausible
that the only true metric performance difference lay in fluoroscopic use.
Nevertheless, fluoroscopic time measurements combined with the number of cine
loops may prove to be a readily exploitable VR-lab metric. Specifically, these two metrics
might serve as indirect measures for radiation exposure and lend themselves to improve a
trainee’s radiation discipline. That is, a trainee could spend time learning procedures in the
VR-Lab and thereby decreasing real world exposure to ionizing radiation to themselves and
patients while simultaneously learning better radiation protection techniques without true
exposure.
Although experienced in general interventional techniques, the CAS procedure was a
novel experience to the expert group. Thus, the expert group’s performance may not truly
represent the abilities of specialists with more CAS experience and should be considered as a
plausible Type II error.
Determining the actual quality of a given interventional performance remains
difficult.81 Furthermore, standardized evaluation is complicated because critical procedural
steps vary between procedure types and different simulators. Parameters such as respect for
tissue, fluoroscopic proficiency, time-motion efficiency, instrument handling/safety and
procedural knowledge have shown promise in assessing interventional skills transfer from the
VR-Lab to the Cath-Lab.82,83 Dayal’s validation also demonstrated that GRS proctor scores
were sensitive in detecting performance differences between different groups both pre- and
post-training.77 Cognitive task analysis of interventional skills currently under the guidance
from the Joint Simulations Task Force (JSTF) of SIR and CIRSE promises to offer more
precise protocols for use in skill assessment in the future.84-86
49
Given that carotid artery stenosis intervention is procedurally difficult, possesses an
extensive learning curve and involves a grave list of potential complications, the role of non-
clinical skills training is of increasing importance.15,81 Indeed, similar arguments were given
when the FDA mandated VR-Lab training in the United States prior to being certified to
perform CAS stenting.87 Thus, further validation studies are of increasing clinical importance
and necessary for the development of standardized curricula.
Three parameters of clinical importance approached statistically significant levels.
Experts outperformed novices with greater tool:vessel ratios and lower residual stenoses,
which translates to wider dilation diameters and better post procedural end organ flow in real
patients. In contrast, novices “outperformed” the experts in placement accuracy by centering
their stents precisely over the carotid lesions. However, in the CAS procedure, the stent is
placed over the lesion with a large portion protruding into the common carotid artery,
meaning that the stent’s center will not lay over a lesion’s true center. Therefore, placement
accuracy seems to be a less meaningful metric compared to the renal artery stenting where
exact stent placement ensures exclusive stenting of only pathological tissue.
SKILLS TRANSFER
Many previous skills transfer studies have focused on time to completion as their
major endpoint.25,45,88 While an expert ought to be able to perform any given case faster than
a novice, time to completion cannot be considered the best indicator of technical skill. The
most important factor is an evaluation of the quality of the finished product.89 Therefore, an
objective method to assess procedural skills was used, i.e. Total score. This method is a
deliberate, standardized complement to the mentoring method wherein an experienced master
judges an apprentice to have become a safe pair of hands. Another benefit of this
methodology is its stability over time. Total score skills assessment is independent of the
laborious necessity of validating new simulator metrics as newer modules and metrics are
developed.
What’s more, many of the experimental attempts to prove that VR-Lab acquired skills
transfer either to another simulator or to the actual OR or Cath-Lab have included control
cohorts which received no training.2,12,13,39,51,52 Interestingly, the evidence from these studies
has not been unanimously supportive for the transfer of VR-Lab skills to real operative
theaters. Thus, at least some of the available evidence does not use an entirely fair
comparison as many have tested one training form against no training at all. Study II
attempted a comparison of the most advantageous training method available, i.e. using the P-
Lab for an approximation of the human Cath-Lab, and produced significant results in Total
scores for both forms of training and therefore supported the skills transfer hypothesis.
In the general statistical model, both training methods improved Total score
significantly and with similar magnitudes for each additional session in either laboratory.
This finding indicates that the two training laboratories were as effective at improving
endovascular skills, yet lack of difference should not be mistaken for equality. Of note, total 50
scores were significantly and consistently higher in the VR-Lab which confirms that the real
catheterization laboratory offers a more challenging venue. An exploitation of this finding
might be the substitution of performance evaluations in the virtual environment for the real
catheterization laboratory. Assuming the predictive value of VR-Lab assessments, attainment
of Total score benchmarks in the VR-Lab might be required prior to allowing access to
perform procedures on patients.67,78,90,91
The data reinforces the maxim that one improves at a skill with specific practice.4,5
Increasing number of P-Lab sessions improved P-Lab Total scores but had no influence on
VR-Lab Total scores. In contrast, previous VR-Lab sessions significantly improved both P-
Lab and VR-Lab Total scores. The evidence indicates that endovascular skills learned on the
Procedicus-VIST™ transfer to the real catheterization laboratory-- as modeled by healthy
porcine anatomy in the P-Lab. Indeed, Chaer et al, using similar performance measurements
as Study II, recently demonstrated significant improvements in the real Cath-Lab in a VR
trained group during proctor supervised iliac artery stenting procedures.83
Why P-Lab sessions failed to improve VR-Lab Total scores is of particular interest.
This failure was assumed to be caused by lack of interest in the participants. In essence, the
“window of opportunity” where the trainees were interested in performing well on the
simulator may have passed, once they had managed the P-Lab session. If true, this may be of
importance when constructing curricula as high motivation is a prerequisite for optimal
training.53,92,93 Therefore, it may be best to initially train in the VR-Lab before moving into
the Cath-Lab to avoid the detrimental effects of boredom.
SUBJECTIVE EVALUATIONS OF THE VR-LAB Subjective evaluations of the VR-Lab were favorable throughout despite observed
differences between students and experts. Students’ responses were generally more positive.
Of note, students expressed increased interest in endovascular specialties and further VR-Lab
training at the end of the experiment suggesting that early experience in a VR-Lab may help
students in making a more informed career choice.94,95 Both groups felt the VR-Lab
experience was realistic compared to their experience in the P-Lab, more effective than
didactics alone and a good teaching instrument which should be mandatory in an
endovascular curriculum thus recognizing that VR is a valuable adjunct to standard training.
The data indicate that most subjects would support a compulsory training curriculum
combining VR-Lab and P-Lab training under the watchful eye of an experienced proctor.
However, porcine experience appeared to be strongly favored over the VR simulation
51
experience. This may be a ripple effect originating either from their low opinion of the
simulator’s realism or the negative influence of software “complications,” i.e. computer
malfunctions, experienced in the VR-Lab. We observed that trainees were keenly interested
in managing OR complications and viewed them as a complement to their training. In
contrast, computer “complications” had no real world value for the trainee and were viewed
as an unwelcome interruption. Subjective and objective data have shown that all participants
left the course with a better grasp of the IAS procedure regardless of which training group
they were assigned. Lastly, in spite of the fact that most felt VR to be a good training tool,
they did not think that it could replace the porcine training laboratory.
All participants felt that proctored instruction was essential and that they had a better
understanding of the procedures after the study. The issue of proctoring is critical. Machines
alone are not an adequate replacement for an experienced teacher. It is certainly true that
most skills laboratories have found repeatedly that faculty involvement during teaching
sessions is a critical element of successfully incorporating VR into procedural training. There
continues to be a vital need for the continued use of the master-apprentice model in
interventional skills education which has served so well in the past. While all groups grasped
the advantages of VR training offered, they acknowledged the continued role of one-on-one
mentoring in their path to skills perfection.
ECONOMICS
Most readers need no introduction to the impact that restrictive budgets have on the
educational goals for institutions of higher learning. The demands of training increasingly
complex surgical and interventional procedures have risen as the available amount of legal
working hours has diminished.54,56-58 Simply, given that two methods of training for
interventional skills produce similar results, institutions should prefer the more economical
method. All health care systems are experiencing financial pressures. This reality mandates
that the training sector seek tools that have the optimal cost-benefit ratio. In other words, we
must attempt to provide the best healthcare and medical training to the greatest number of
individuals using finite resources. This philosophy is extant whether the funds available are
coming from public or private sources.
Stents in the P-Lab are consumed during each procedure since a deployed stent cannot
be practically retrieved. Even at volume discounted prices, this accounted for 83% of the total
cost in the P-Lab. In contrast, the interventional stents used in the VR-Lab can be reused
since they are “virtually” placed and constituted only 6% of the total VR-Lab costs assuming 52
a simulator purchase, or 20% under the rental analysis. Naturally, training facilities could use
older equipment from national Cath-Labs which had expired dates to offset this cost but this
would lead to a course which trained novices with old, if not obsolete, techniques. Thus, the
cost for an institution for a modern P-Lab stenting course would vary in proportion with the
amount of stenting planned.
The purchase price of a Procedicus-VIST™ spread over a five year period as the
annual depreciation value of €30,000 represented the brunt of the VR-Lab at 87% compared
to the annual rental of the P-Lab at €4,140, i.e. 9% of the budget. However, this calculation
assumes a residual value of the VIST™ to be €50,000 and institutional ownership of a
portable VR-Lab. Perhaps of greater importance, it might be possible to upgrade the
simulator with newer software packages or hardware thereby extending its life cycle beyond
the five year estimate. If true, the annual depreciation value would then be modified to the
new life cycle end making the true annual cost smaller than the current estimate used in the
analysis. Additionally, the simulator could be rented out or used as part of a free-standing
endovascular curriculum given by the purchasing institution thereby functioning as a source
of potential income.
While the cost ratio from VR-Lab purchase vs. P-Lab rental analysis found the VR-
Lab to be 0.75 the P-Lab course costs for the first year, the VR-Lab/P-Lab ratio in the
sensitivity analysis ranged from 0.25 to 2.22, with six of the nine possible outcomes favoring
the VR-Lab, Table 14. However, two of the three ratios which favored the P-Lab over the
VR-Lab assumed a price increase for the VIST™. Although technical refinements and
improvements of the simulator hardware and software performance may to some extent add
to the cost, such a scenario can be considered highly unlikely as information technology (IT)
based industries depend upon delivering more powerful computers at lower costs at regular
intervals. The remaining favorable P-Lab ratio assumes a sudden decrease in the price of the
P-Lab rental fee which can be considered to be unlikely. Therefore, the existing ratio or those
which include a stable or falling VR-Lab estimate are the most plausible findings in the
sensitivity analysis. The VR-Lab rental vs. P-Lab rental based analysis cost ratios were
unanimously in favor of the VR-lab at all cost estimate fluctuations. However, as access to
rental simulators was assumed to be the limiting factor for most institutions, the VR-Lab
purchase vs. P-Lab rental was deemed to be the financial analysis with the most external
validity, i.e. general applicability.
53
Finally, learning a skill and perfecting it are two different phases in the pursuit of
mastery of manual skills training.30 Repetition is an essential part of mastery, but—
presently—trainees normally use patients to perfect their skills. The natural extension of an
economic analysis then becomes how many courses are required to reach a certain skill level
and how do their efficacies compare, i.e. does one alternative offer quicker mastery compared
to the other and, if so, how does this affect the cost ratios. Thus, not only must introductory
courses be considered, but also the number of courses needed to attain benchmark levels prior
to practicing in the human catheterization laboratory. In addition, refresher courses may be
needed to maintain skills for low frequency procedures in clinical practice. With the current
level of evidence regarding efficacy and curricula, one cannot make any meaningful
economic calculations based upon required course numbers as that number is unknown.
Rapid advances in computer sciences over the last decade have heralded phenomenal
technology such as virtual reality simulators capable of reproducing operative experiences
outside of the OR. Although we may still be in the embryonic stages of VR development, its
future appears bright. Therefore, investment in a technology that possesses as many intrinsic
advantages as VR does, seems to be a wise strategy for the future. However, acceptance of
the VR-lab as a venue for skills acquisition within the scientific community awaits further
evidence demonstrating the transfer of skills from VR to the Cath-Lab.32,45-47,50,83 As rational
as this requirement seems, the same demands were never placed upon current non-clinical
training using research or experimental animals. Remarkably, no literature supporting
endovascular skills transfer from the live animal model to the OR could be found. Thus, as
demonstrated by the lack of evidence regarding both labs, we still know very little about what
type of training helps in the real world, Table 1.
Evidence supporting skills transfer from endoscopic and laparoscopic VR-Labs to the
OR exists.12,39,45,65,91 However, extrapolation of those results to endovascular techniques was
deemed inappropriate because many of these studies compared the VR-Lab group to non-
trained controls and, more importantly, lacked an endovascular focus. One study
demonstrated transfer of VR-Lab acquired skills to the P-Lab.96 No studies were found to
have compared the efficacies of the P-Lab to the VR-Lab in acquiring or perfecting
endovascular skills in the human catheterization laboratory.
54
Despite the lack of confirmative data regarding the clinical efficacy of both P-lab and
VR-lab endovascular skills training, the results of the financial calculations raise an important
ethical question. If skills training in the VR-lab is much less expensive than training in the P-
lab, and if there is no proven superiority of using an animal lab, can skills training on animals
be supported from an ethical perspective? In order to substantiate continued use of live
animals, the outcome of skills training in the P-lab should be significantly more efficient than
training in a VR-lab. Although further refinements of the VR-lab modules may be necessary
to fully compete with P-lab training, the potential for technical development of the VR-lab
suggests that the need for live animals may be abolished in the future.
ETHICS
Ziv et al produced the most recent article to bring simulations and medical ethics
together in 2003.97 The paper acknowledged that medical training, at some point, must use
live patients, but that computer simulation based technology could be a valuable tool in
mitigating the ethical tensions and practical dilemmas these first steps entail. Their proposed
ethical framework focused on best standards of care and training, management of medical
errors, respect for patient safety and autonomy and the responsible allocation of resources.
Thus, the implementation of increased VR-Lab endovascular training complies with that
framework and succeeds in fulfilling our ethical responsibility as researchers and educators.
Although not focused solely upon VR and ethical considerations, Gruber and Hartung
pointed out that the validity of some animal research is often unknown either through lack of
will or ignorance.98 Furthermore, because the data produced using animals is commonly
statistically under-powered, it is impossible to draw reliable conclusions from the observed
outcomes. The parallel between Gruber and Hartung’s observations and the uncertainty
surrounding P-Lab skills transfer efficacy are not difficult to recognize. Even though the P-
Lab mimics the OR and involves the use of real interventional equipment to develop skills, it
does so on non-human species with healthy anatomy and produces unknown results in the
trainee. Studies need to be performed to quantitatively measure the improvements that have,
thus far, been assumed by training on pigs. The same arguments hold true for the VR-Lab. As
the experts in this field have recently pointed out, “Intuition is not evidence.”50 Yet, neither
does lack of evidence mean that something is untrue, merely unproven.
On the other hand, the possibility of training a technique incorrectly and/or building
false confidence exists when using an untried method. What if one or both of the labs teach,
unknowingly, a skill that will be harmful to cath-lab performance? One cannot assume that
these training forms offer only benefit; it must be proven. Also, while mechanical skills can
be taught given the proper setting, most experienced interventional radiologist will agree that
clinical judgment and situational awareness are the hardest to teach. Specifically,
overstepping one’s ability due to overconfidence and ignoring common problems due to
ignorance seems to plague beginners.
55
Despite the fact that the general public’s emphasis on ethical responsibility commonly
outweighs economical issues, pragmatic use of finances is necessary. Gruber and Fitzpatrick
state that ethical considerations alone will do little to bring about change.98 Furthermore, they
remind us of our responsibility to ask only important questions and to design experiments
with ethics in mind which minimize suffering.98,99
According to Machan, animals lack rights because they do not have the faculty to
make moral choices.100 Thus, their unaccountable actions are a consequence of instincts
rather than conscious deliberation between what is right or wrong. He reminds us that
humans are also a part of nature and, as such, make use of the animals below us to survive
and thrive, as do all of nature’s creatures. However, even though the lesser animals lack such
moral agency, we are bound by ethics to treat them with consideration. In essence, we ought
to be thinking of a better way to get the experimental data humanity needs while including
the three R’s, i.e. 1) Replacement 2) Reduction and 3) Refinement, at each step.
Thus, the main argument for P-Lab training is to increase patient safety during the
steepest part of the learning curve. Improvement in endovascular skills does not appear to be
part of the argument. In fact, the efficacy of P-Lab training on human catheterization
laboratory is unproven placing it on a level of uncertainty equal to the VR-Lab. Therefore,
two alternative methods of training endovascular novices are available, which offer the
benefit of early learning curve safety to patients. The choice of which alternative to use
remains unanswered yet the default remains the P-Lab in countries where it is available.
56
Thus far, we have dealt with skills training removed from the real catheterization
laboratory in order to increase patient safety and save costs. Yet our most common training
venue remains the human cath-lab. While skills transfer from the human cath-lab is a moot
point, the ethics of using patients for skill acquisition and practice remains problematic. Is it
ethical to include humans in our training? The benefit to the patient—and potential harm—
must be weighed against the student’ benefit as they acquire experience. Even though the use
of research animals raises ethical concerns as discussed previously, should “skills training”
be considered research or merely a form of consumption? Research animals are used in the P-
Lab to gain practical experience removed from the human laboratory, not to increase
scientific knowledge. We train non-clinically that we might spare fellow humans harm
caused by our lack of experience. Yet, can we make the decision to expend another sentient
being without really knowing if it serves its stated purpose? The addition of validated training
method(s) would certainly be welcomed by fully informed, consenting patients prior to
becoming a fellow’s first stenting procedure. Furthermore, an adequately trained fellow
might enter the human laboratory at a higher point on their learning curve with a decreased
likelihood of errors and faster procedure times, both of which lead to cost reduction.65
57
STUDY LIMITATIONS
First, the sizes of the studies are a limiting factor. In the validation studies I and IV,
we were unable to recruit sufficient numbers of “unexposed” novices and interventionalists to
substantiate the lack of evidence. However, post-hoc analysis for the two key metrics of
residual stenosis and placement accuracy revealed that the numbers needed were not possible
to reach within Sweden. What’s more, the use of the simulator as a proficiency assessment
instrument is, at best, an “off-label” use as it was designed to educate not regulate. The final
approval of a novice’s clinical proficiency ought to lie in the hand of an endovascular expert
with years of clinical experience, and not in the hands a software programmer.
Study II was also small due to the logistical and financial limitations of the
laboratories used. Nevertheless, the study design did prove valuable in detecting a
transferability of virtually generated skills. The measured skill improvements may not
represent the maximized training benefit of either method. A fairer comparison would have
included longer training periods in both laboratories. The small but significant gains observed
may lack clinical relevance. In the general statistical model, either training form improved
Total score by only 6.8% of the total maximum possible per session. Lastly, the significance
level for the critical result, i.e. VR-Lab sessions improved P-Lab performance, lay just within
the acceptable p = 0.05 level. Stronger significance would have enhanced the trustworthiness
of our findings.
Any financial analysis is rift with assumptions, which are presumed to be correct but
are often only educated guesses based on local and historical financial data. Thus, despite
including sensitivity analysis for both the VR-Lab purchase/rental vs. P-Lab rental, real
economic data could vary greatly not only from country to country but also between
institutions within the same metropolis. Therefore, the true external validity of the findings,
although likely to be accurate in some situations, could be unrealistic in others.
Finally, these projects have included only one type of endovascular simulator. Several
are available on the open market, but our laboratory only had access to one due to funding
issues. The conclusions drawn from the data relate only to this particular simulator and not to
VR simulations in general. Better methodology would have included each of the existing
simulators in their own experimental group to detect possible differences between the models
as has been done for other specialties.101
59
FUTURE RESEARCH
The true litmus test of this simulator’s usefulness will lie in its ability or failure to
significantly improve human catheterization laboratory performances vis-à-vis improved
objective performance scores significantly related to VR training. However, definitive proof
for the reduction of intraoperative errors would also bolster support for the incorporation of
skills training in the VR-Lab. To that end, a human transfer study is in progress at the
Interventional Radiology Department, Kuopio University, Finland using a similar clinical
assessment scale as the Total Score used in Study II, renamed the Objective Assessment of
Scandinavian Interventional Skills (OASIS). Results are expected by the end of 2007.
Furthermore, the rapid advances in computer technology all but guarantees periodic
software updates to the metrics of this and other simulator systems. Naturally, once a new set
of metrics are introduced, they too will need to be systematically validated. As most
validation studies consist of relatively small n-values—our studies included—confidently
claiming or denouncing construct validity will continue to be based upon lower powered
studies. In order to automate the collection and analysis of larger data sets, a kiosk database
program has been developed which could be delivered to high volume VR training centers
worldwide such as Abbott Vascular’s Crossroads training institute in Diegem, Belgium.
Software-version-specific metric data can, in this fashion, be collected in conjunction with
demographic/IR experience data from the different locations. Periodic downloading to a
central validation authority, such as the Joint Simulation Task Force (JSTF) would greatly
speed the validation process and increase our confidence in the use of VR technology.
Yet another application of simulation technology is the ability to perform “mission
rehearsals”, termed “procedure rehearsal” within medicine, for critical procedures such as
neuroradiological interventions. In collaboration with Anthony Gallagher, we have initiated a
live case procedure rehearsal study scheduled for May 2007. The participants include
neuroradiologists from Turkey and four patients, two of which will have had their CT scans
converted into patient specific VR modules on the Procedicus-VIST™. The potential benefits
of being able to perform a clinical case in the Cath-Lab which one has practiced to perfection
in the virtual environment may prove to be of increasing clinical importance in the future.
Lastly, the previously mentioned JSTF has launched a detailed task analysis of many
IR procedures. The major aim of this collaboration is to dissect complex procedures down to
their basic steps with the help of cognitive behavioral psychologists. These steps will then be 61
weighted according to their importance to the end quality of the intervention in question. In
this way, a higher quality skills assessment scale is being devised to produce clinically
relevant performance measurements and potential new metric parameters. The results of this
effort will finally give the endovascular community their gold standard. More importantly,
the information could be given openly to the industry to stimulate the development of
clinically significant metrics for future VR simulators. In this way, the medical community
could step forward to take a proactive role in developing future medical simulators versus
waiting for the industry to independently create them in the absence of guidance as to what is
important to the end users.
Although currently useful for endovascular training, further improvements in the VR-Lab are needed for this training form to reach its full potential as a true full-scale interventional simulator. Namely, the inclusion of interventional and clinical complications such as sub-intimal catheterizations, iatrogenic thromboembolism to end organs, coronary arrythmias/infarction, apoplexia, acute reactions to contrast medium and vessel puncture complications.. Furthermore, femoral artery puncture, i.e. the very start of a procedure, constitutes a substantial obstacle for beginners. There is always a risk for dissection, embolus, thrombosis, development of hematoma and pseudoanerysms after an arterial puncture. Likewise, arteriotomy closure, using various closure device instruments presents another application which lends itself to virtual reality training in order to avoid or minimize the incidence of post-operative complications such as pseudo aneurysms, inguinal hematomas and fistula formations. In order for such developments to take place, the medical community should initiate or increase their contact and cooperation with the industry.
62
CONCLUSIONS
I. With the exception of total procedure times, the renal artery stenosis modules failed to
demonstrate construct validity. A) Despite demonstrating face validity and some non-
significant absolute performance metric differences between groups, the virtual reality
simulator was unable to stratify virtual interventional performances based upon
experience. B) Current renal module metric parameters were not capable of
performance assessment outside the catheterization laboratory. C) The VR-Lab was
generally approved as a pedagogic tool based upon the subjective data.
II. Porcine and virtual reality laboratory training appears to produce comparable training
results in the endovascular novice as measured in the porcine laboratory. A) Virtual
reality simulation training was found to be as effective as the porcine laboratory in
improving P-Lab iliac artery stenting performance. B) Skills learned in virtual reality
using the iliac artery modules may transfer to the catheterization laboratory as simulated
by healthy porcine anatomy. C) While both laboratories were subjectively approved for
endovascular skills education, the P-Lab was apparently favored by trainees based upon
subjective data.
III. An economical analysis of using virtual reality simulations versus the porcine
laboratory to train endovascular skills was performed. A) The purchase of a simulator
cost less compared to renting the P-Lab in the short and long term perspective. B)
Rental of a VR-Lab was found to be considerably less expensive than renting the P-Lab
to meet national training demands, although locations without adequate rental access
would make such an implementation difficult.
IV. With the exception of procedure and fluoroscopic times, the carotid artery modules
failed to demonstrate construct validity. A) Despite demonstrating face validity, the
simulator metrics were unable to stratify virtual interventional performances based upon
prior endovascular experience. B) The Procedicus-VIST™ metrics did not work as an
assess tools for endovascular skills outside of the catheterization laboratory. C) The
VR-Lab was subjectively approved as a pedagogic tool.
63
SVENSK SAMMANFATTNING
BAKGRUND: Interventionell radiologi av idag innebär diagnostik och behandling av
sjukdomar med hjälp av perkutana minimalinvasiva, oftast kateterstyrda procedurer med stöd
av bildgivande radiologiska metoder (datortomografi, magnetkamera, röntgengenomlysning,
angiografi, ultraljud). Endovaskulär radiologi är en förhållandevis ung gren av medicinen och
etablerades först av pionjärer på 1960-talet. Den första perkutana transluminala
kateterbehandlingen rapporterades av Dotter i femoro-popliteala kärlområdet i Circulation
1964. Tekniken vidareutvecklades av Gruntzig till att innefatta en angioplastikballong och via
Palmaz introducerades på 80-talet stentbehandlingar av blodkärl. Liksom inom kirurgin har
interventionell radiologi anammat samma utbildningskoncept, dvs. att överföringen av
kunskap ska ske i ett mästar-lärlingsförhållande där röntgenrummet med angiografiutrustning
utgör klassrummet. Basal endovaskulär teknik har därför utförts på människa direkt, även om
detta inte varit direkt utsagt.
Det finns många faktorer som talar för att den basala träningen av interventionell
radiologi ska flyttas ut från den direkta patientkontakten i operationsrummet: För det första
har en rapport från USA (Kohn, Corrigan, Donaldsson, 1999) påvisat att mellan 44000 och
98000 individer per år avlider till följd av misstag inom sjukvården. Fel eller misstag inom
sjukvården kan indelas i systemfel och i individuella fel. Analys och tillfälle till reflektion
möjliggör att vi kan lära av våra misstag. Det måste då finnas säkerhetssystem som objektivt
kan utvärderas. Kvalitetssäkring av utbildning, vidareutbildning och kompetensbevarande
utbildning är nödvändigt. Ahlberg (Thesis Ahlberg, 2005) har kritiserat den äldre mästar-
lärlingfunktionen bl.a. därför att den är svår att utvärdera.
För det andra är bildstyrd intervention där operatören vägleds av en tvådimensionell
monitorbild av en patients inre komplex. Metoden kan anses vara relativt sett svårare att lära
än öppen kirurgi. Höga krav ställs på visuell och spatial förmåga. Den ökade komplexiteten
innebär också en ökad risk för komplikationer.
För det tredje har interventionell radiologi de senaste åren fått en mycket stor och
ökande betydelse som alternativ och komplement till sedvanlig kirurgisk behandling av
sjukdomar i kärlsystemet. Genom dess minimalt invasiva karaktär har dessa interventionella
metoder lett till att patientgrupper som tidigare inte kunnat behandlas på grund av risk för
komplikationer, nu kan åtgärdas med framgång och med betydligt mindre risker.
65
Komplicerade nya endovaskulära tekniker har införts, t.ex. endovaskulär aortareparation,
EVAR. De kärlproteser som finns kräver utbildning i handhavande och certifiering och
alltmer komplicerad teknologi är på väg. Risken för komplikationer ökar med ökande
komplexitet.
För det fjärde finns det i all färdighetsutövande en inlärningskurva innan man uppnått
en godtagbar standard. De flesta svårigheter och komplikationer som inträffar, sker tidigt i
inlärningsfasen men bidrar också till kunskapsinhämtande. Att flytta träning från angiosalen
till en simulator skulle kunna förbättra säkerheten, då de initiala momenten kan övas till dess
de integreras som en färdighet (jämför att lära sig cykla, köra bil). Träning av medicinska
färdigheter samt träning av kommunikation, ledarskap och samarbete i simulatormiljö bör
leda till en ökad patientäkerhet. Misstag kan minimeras om medicinska färdigheter och
kompetenser kan öka i simulatormiljö.
För det femte är en grundläggande färdighet i punktionsteknik och invasiv angiografi
en del i att utvecklas till en duktig interventionalist. Antalet utförda angiografier som krävs
innan en operatör kan utföra en angioplastik under handledning är en bedömningsfråga. De
enkla diagnostiska angiografierna har nu minskat dramatiskt och möjligheterna för unga
läkare att få en basal färdighet har därmed minskat. Alternativa metoder är träning på
djur/kadaver för att skona patienter från otränad personal eller att öva med medicinska
simulatorer.
Syftet med denna avhandling har varit:
att undersöka betydelsen av simuleringsteknologi med avseende på validering av en
simulator för kateterburen diagnostik och intervention
att jämföra endovaskulär träning på simulatorn med träning på djur
att genomföra en ekonomisk analys av kostnaderna för simulatorträning med
kostnaderna för träning på gris.
66
I DELARBETE I var målsättningen att studera om simulatorn kunde diskriminera
mellan erfarna radiologer (n=8) och läkarstudenter (n=8) på termin VIII, med hjälp av de
metriska resultat som simulatorn levererar. Alla deltagarna fick en kort introduktion i
endovaskulär teknik och i handhavandet av simulatorn. Därefter fick alla deltagarna på
förmiddagen arbeta med sex olika patientfall med njurartärförträngning. På eftermiddagen
registrerades deltagarnas resultat i en ny omgång på alla sex fallen. Det visade sig att
studenterna använde längre genomlysningstid (mer röntgenstrålning) än erfarna läkare, men i
övrigt skiljde sig inte resultaten signifikant åt. Någon värdering av deltagarnas färdigheter av
en erfaren radiolog gjordes inte i denna studie. Både studenter och erfarna radiologer
bedömde att simulatorn är ett viktigt pedagogiskt instrument och att den är en god spegling
av verkligheten (face validity).
I DELARBETE II deltog tolv läkare med varierande erfarenhet i interventionell
radiologi i en tvådagars kurs för att utveckla sin teknik i endovaskulär stentning av
bäckenartärer. Deltagarna indelades i 4 grupper. Med hjälp av en utsänd enkät lottades
deltagarna via ett statistiskt förfarande så att kunskapsnivån i de 4 grupperna skulle kunna
vara så lika som möjligt med avseende på erfarenhet. Den första gruppen fick enbart träna
endovaskulär teknik på gris, den fjärde gruppen enbart på en medicinsk simulator under de
två kursdagarna. Mellangrupperna fick träna på gris första dagen och simulator dag två.
Deltagarnas basala kunskapsnivåer värderades av särskilt utsedda erfarna lärare både på gris
och simulator vid fyra tillfällen, d.v.s. på morgonen och vid dagens slut båda kursdagarna. Ett
särskilt utvärderingsprotokoll modifierat från Dr Reznick, Professor, Toronto, Canada och Dr
Beard, Consultant Vascular Surgeon, Leicester, England användes. Videoinspelning av
deltagarnas prestationer utvärderades separat av två mycket erfarna specialister i ett senare
skede. Simulatorträning och gristräning visade sig båda vara effektiva i att förbättra den
endovaskulära tekniken. Den var en god samstämmighet mellan videobedömning och
bedömning på plats av deltagarnas insatser. Träning i simulatormiljö medförde att deltagarna
presterade bättre vid den därpå följande träningen på gris (som ska efterlikna
patientsituationen). Båda träningsmetoderna ansågs av kursdeltagarna vara goda miljöer för
träning.
I DELARBETE III analyserades de ekonomiska kostnaderna i studie II. Två scenarier
analyserades: Först jämfördes kostnaden att köpa en simulator med att hyra ett
gristräningslaboratorium under fem år. Denna analys ansågs mest realistisk, då många
institutioner idag har ett djurlaboratorium, men få institutioner äger en simulator. Sist
jämfördes också kostnader för att hyra simulatorn med att hyra ett djurlaboratorium under
samma period, d.v.s. fem år. De faktiska kostnaderna för vårt försök i delarbete 2, liksom de
båda beskrivna scenarierna visade att en medicinsk simulator är betydligt mer ekonomiskt
fördelaktig jämfört med träning på gris. Sannolikt blir simulatorerna allt mer sofistikerade
och billigare i framtiden. Etiskt framstår medicinska simulatorer som ett självklart
förstahandsval för träning av endovaskulära procedurer.
67
I DELARBETE IV deltog 16 läkare med varierande erfarenhet av interventionell
radiologi och 16 läkarstudenter från den avslutande delen av sin utbildning. En komplicerad
endovaskulär procedur, stentning med angioplastik av arteria carotis interna (halspulsådern)
med filterskydd, introducerades för deltagarna. Ingen av deltagarna hade utfört en sådan
procedur på människa. I denna studiedesign fick kursdeltagarna en kort introduktion av
proceduren och i handhavandet av simulatorn. Deltagarna fick sedan öva på ett fall, därefter
värderade simulatorn resultaten på nästa fall. Det visade sig att den totala behandlingstiden
och röntgengenomlysningstiden var signifikant kortare för de erfarna läkarna jämfört med
studenterna. De båda grupperna angav att simulatorn är ett bra pedagogiskt verktyg och också
efterliknar verkligheten väl.
SAMMANFATTNING: Våra resultat visar att en endovaskulär simulator är ett
betydelsefullt pedagogiskt instrument och också en god spegelbild av den endovaskulära
miljön, samt ett ekonomiskt fördelaktigt alternativ till träning på djur. Simulatorer kan dock
inte idag, med de återkopplingsinstrument (metrics) som finns tillgängliga, visa skillnaden
mellan erfarna och oerfarna operatörer, mer än i tidshänseende och i användandet av
röntgengenomlysning.
Framtiden: Man bör i framtida studier analysera vilka moment i en endovaskulär procedur
som är viktiga att lära sig och hur komplikationer och bra respektive dålig teknik skall
defineras. Denna kunskap bör sedan implementeras i simulatorernas återkopplingssystem
(metrics). Kanske kan då eleverna träna ensamma på simulatorerna tills de uppnår en viss
färdighet. För att få ut så mycket som möjligt av träning med en endovaskulär simulator bör
man också integrera flera utbildningsaspekter i ett curriculum. Träningen bör ses ur ett
helhetsperspektiv. Att optimera utbildningen avseende självstudier, föreläsningar om
proceduren, träning i simulator med och utan handledning och på sikt kunna definiera en
certifieringsnivå för utförande av procedurer på patienter är utmaningar för framtida
forskning.
68
ACKNOWLEDGEMENTS
Richard Reznick, MD, M.Ed., FRCSC, FACS, Professor and Chair, Department of
Surgery, University of Toronto for his enormous contributions and support from
abroad.
Ted Lystig, Ph.D., Biostatistician, AstraZeneca; a selfless and congenial biostatistician of
unrivaled caliber.
Jonathan Beard, M.B.B.S., M.Ed., FRCS, Department of Vascular Surgery, Sheffield,
United Kingdom whose generosity and encouragement has been priceless.
Hans Klingenstierna, M.D., Department of Interventional Radiology, Södra Älvsborg
Sjukhus, Boras, Sweden for his directness, brilliant work ethic and almost permanent
smile.
Jan Göthlin, M.D., Ph.D., Professor Emeritus, Department of Radiology, The Sahlgrenska
Academy at Göteborg University, Sweden for inspiring me to dig deeper and fight the
good fight.
Mikael Hellström, MD, PhD, Professor and Chair, Department of Radiology, The
Sahlgrenska Academy at Göteborg University, Sweden for allowing this dissertation
to be undertaken within his department and his tireless support through the highs and
lows.
Finally, I want to express my gratitude for my academic supervisor/mentor/trouble-
maker, Lars Lönn. Without your contagious energy and inspiration, I’d be much better rested
(yet unenlightened). Deepest thanks.
Academic Supervisor, Lars Lönn, M.D., Ph.D., Associate Professor, Department of
Radiology, Sahlgrenska University Hospital, The Sahlgrenska Academy at Göteborg
University, Sweden and Department of Radiology, Rigshospitalet, Copenhagen,
Denmark
69
Department of Radiology, Sahlgrenska University Hospital, in particular the Uro-Gastro
Section without whom, Paper II would have never been possible.
Department of Interventional Radiology, Rigshospitalet, Copenhagen Denmark, for the
opportunity to gain some clinical experience under their expert tutelage.
Robert Emilsson, Photographer, The Sahlgrenska Academy at Göteborg University,
Sweden for expert videographic contributions.
Professor Mats Sandberg, Ph.D., Department of Biochemistry, The Sahlgrenska Academy
at Göteborg University, Sweden for introducing me to scientific research.
Professor Keiko Funa, M.D., Ph.D., Department of Biochemistry and Cellbiology, The
Sahlgrenska Academy at Göteborg University, Sweden for allowing me to test the
waters of research in her prestigious laboratory.
Erney Mattson, M.D., Ph.D., The Wallenberg Laboratory, The Sahlgrenska Academy at
Göteborg University, Sweden for his insightful and beneficial suggestions.
Dan Lukes, M.D., Ph.D., The Wallenberg Laboratory, The Sahlgrenska Academy at
Göteborg University, Sweden for his kind words of encouragement and guidance.
Lena Björneld, Department of Radiology, The Sahlgrenska Academy at Göteborg University,
Sweden for her expert editorial assistance.
Ingela Björholt, PhD., Nordic Health Economic Research AB, Sweden for help in sorting out
the details of the financial analysis.
Lotta Robertsson, The Sahlgrenska Academy at Göteborg University, Sweden
for always having patience and a few kind words to spare.
Mentice Medical Simulations, Gothenburg, Sweden for magnitudes of cooperation and their
desire to make simulator technology mainstream.
Experimental Biomedical Institute especially Eva Oscarsson and her team.
Bard Nordic AB, Martin Bengtsson and Lotta Schrewelius.
Cordis AB, Johnson and Johnson, Daniel Karlsson and Henrik Nyman.
General Electric AB, Sweden
Philips AB, Sweden
Cardiovascular and Interventional Radiological Society of Europe, CIRSE, Educational
Research Grant 2005
Göteborgs Läkarsällskap for graciously funding portions of this work.
Svensk Förening för Medicinsk Radiologi
70
I would also like to express my gratitude to the following for their unconditional and
untiring support:
Sidsel Berry
Maw and Paw
Inge and Berit Danielsson
Jimmy and Helle Witved
Martin Carlwe, M.D.
Shahin Mohseni, M.D.
Malin Lönn, Ph.D.
Thanks for putting up with me,
Max
71
REFERENCES
1. Lobato AC, Rodriguez-Lopez J, Diethrich EB. Learning curve for endovascular
abdominal aortic aneurysm repair: evaluation of a 277-patient single-center
experience. J Endovasc Ther 2002;9(3):262-8.
2. Ahlberg G, Hultcrantz R, Jaramillo E, Lindblom A, Arvidsson D. Virtual reality
colonoscopy simulation: a compulsory practice for the future colonoscopist?
Endoscopy 2005;37(12):1198-204.
3. Piechaud PT, Pansadoro A. Transfer of skills from the experimental model to the
patients. Curr Urol Rep 2006;7(2):96-9.
4. Tassios PS, Ladas SD, Grammenos I, Demertzis K, Raptis SA. Acquisition of
competence in colonoscopy: the learning curve of trainees. Endoscopy
1999;31(9):702-6.
5. Parry BR, Williams SM. Competency and the colonoscopist: a learning curve. Aust N
Z J Surg 1991;61(6):419-22.
6. Champion HR, Gallagher AG. Surgical simulation - a 'good idea whose time has
come'. Br J Surg 2003;90(7):767-8.
7. Vozenilek J, Huff JS, Reznek M, Gordon JA. See one, do one, teach one: advanced
technology in medical education. Acad Emerg Med 2004;11(11):1149-54.
8. Ziv A, Ben-David S, Ziv M. Simulation based medical education: an opportunity to
learn from errors. Med Teach 2005;27(3):193-9.
9. Gallagher AG, Cates CU. Virtual reality training for the operating room and cardiac
catheterisation laboratory. Lancet 2004;364(9444):1538-40.
10. Madan AK, Frantzides CT, Sasso LM. Laparoscopic baseline ability assessment by
virtual reality. J Laparoendosc Adv Surg Tech A 2005;15(1):13-7.
11. Madan AK, Frantzides CT, Tebbit C, Quiros RM. Participants' opinions of
laparoscopic training devices after a basic laparoscopic training course. Am J
Surg 2005;189(6):758-61.
12. Youngblood PL, Srivastava S, Curet M, Heinrichs WL, Dev P, Wren SM.
Comparison of training on two laparoscopic simulators and assessment of skills
transfer to surgical performance. J Am Coll Surg 2005;200(4):546-51.
73
13. Hyltander A, Liljegren E, Rhodin PH, Lonroth H. The transfer of basic skills
learned in a laparoscopic simulator to the operating room. Surg Endosc
2002;16(9):1324-8.
14. Korndorffer JR, Jr., Dunne JB, Sierra R, et al. Simulator training for laparoscopic
suturing using performance goals translates to the operating room. Transfer of
skills from the experimental model to the patients Training in laparoscopic
urology. J Am Coll Surg 2005;201(1):23-9.
15. Lin PH, Bush RL, Peden EK, et al. Carotid artery stenting with neuroprotection:
assessing the learning curve and treatment outcome. Am J Surg 2005;190(6):850-
7.
16. Eversbusch A, Grantcharov TP. Learning curves and impact of psychomotor
training on performance in simulated colonoscopy: a randomized trial using a
virtual reality endoscopy trainer. Surg Endosc 2004;18(10):1514-8.
17. Hammond I, Karthigasu K. Training, assessment and competency in gynaecologic
surgery. Best Pract Res Clin Obstet Gynaecol 2005.
18. Schijven MP, Berlage JT, Jakimowicz JJ. Minimal-access surgery training in the
Netherlands: a survey among residents-in-training for general surgery. Surg
Endosc 2004;18(12):1805-14.
19. Cosman PH, Cregan PC, Martin CJ, Cartmill JA. Virtual reality simulators:
Current status in acquisition and assessment of surgical skills. ANZ Journal of
Surgery 2002;72(1):30-34.
20. Madan AK, Frantzides CT, Shervin N, Tebbit CL. Assessment of individual hand
performance in box trainers compared to virtual reality trainers. Am Surg
2003;69(12):1112-4.
21. Mantovani F, Castelnuovo G, Gaggioli A, Riva G. Virtual reality training for health-
care professionals. Cyberpsychol Behav 2003;6(4):389-95.
22. Gallagher AG, McClure N, McGuigan J, Crothers I, Browning J. Virtual reality
training in laparoscopic surgery: a preliminary assessment of minimally invasive
surgical trainer virtual reality (MIST VR). Endoscopy 1999;31(4):310-3.
23. Colt HG, Crawford SW, Galbraith O, 3rd. Virtual reality bronchoscopy simulation:
a revolution in procedural training. Chest 2001;120(4):1333-9.
24. Gor M, McCloy R, Stone R, Smith A. Virtual reality laparoscopic simulator for
assessment in gynaecology. Bjog 2003;110(2):181-7.
74
25. Carter FJ, Schijven MP, Aggarwal R, et al. Consensus guidelines for validation of
virtual reality surgical simulators. Surg Endosc 2005;19(12):1523-32.
26. Piaget J. The stages of the intellectual development of the child. Bull Menninger Clin
1962;26:120-8.
27. Jandering A. Virtual Skill Training: A Learning Theory Approach. Masters Thesis:
Chalmers University of Technology, 2006.
28. Vroom VH. Work and motivation. New York: John Wiley, 1964.
29. Barnes RW. Surgical handicraft: teaching and learning surgical skills. Am J Surg
1987;153(5):422-7.
30. Reznick RK, MacRae H. Teaching surgical skills--changes in the wind. N Engl J
Med 2006;355(25):2664-9.
31. Dankelman J, Wentink M, Grimbergen CA, Stassen HG, Reekers J. Does virtual
reality training make sense in interventional radiology? Training skill-, rule- and
knowledge-based behavior. Cardiovasc Intervent Radiol 2004;27(5):417-21.
32. Gallagher AG, Ritter EM, Champion H, et al. Virtual reality simulation for the
operating room: proficiency-based training as a paradigm shift in surgical skills
training. Ann Surg 2005;241(2):364-72.
33. Sherman V, Feldman LS, Stanbridge D, Kazmi R, Fried GM. Assessing the learning
curve for the acquisition of laparoscopic skills on a virtual reality simulator.
Surg Endosc 2005;19(5):678-82.
34. Brunner WC, Korndorffer JR, Jr., Sierra R, et al. Laparoscopic virtual reality
training: are 30 repetitions enough? J Surg Res 2004;122(2):150-6.
35. Gallagher AG, Satava RM. Virtual reality as a metric for the assessment of
laparoscopic psychomotor skills. Learning curves and reliability measures. Surg
Endosc 2002;16(12):1746-52.
36. Gallagher AG, Lederman AB, McGlade K, Satava RM, Smith CD. Discriminative
validity of the Minimally Invasive Surgical Trainer in Virtual Reality (MIST-
VR) using criteria levels based on expert performance. Surg Endosc
2004;18(4):660-5.
37. Gallagher AG, Richie K, McClure N, McGuigan J. Objective psychomotor skills
assessment of experienced, junior, and novice laparoscopists with virtual reality.
World J Surg 2001;25(11):1478-83.
75
38. Martin JA, Regehr G, Reznick R, et al. Objective structured assessment of technical
skill (OSATS) for surgical residents. Br J Surg 1997;84(2):273-8.
39. Grantcharov TP, Kristiansen VB, Bendix J, Bardram L, Rosenberg J, Funch-Jensen
P. Randomized clinical trial of virtual reality simulation for laparoscopic skills
training. Br J Surg 2004;91(2):146-50.
40. Faulkner H, Regehr G, Martin J, Reznick R. Validation of an objective structured
assessment of technical skill for surgical residents. Acad Med 1996;71(12):1363-5.
41. Beard JD, Jolly BC, Newble DI, Thomas WE, Donnelly J, Southgate LJ. Assessing
the technical skills of surgical trainees. Br J Surg 2005;92(6):778-82.
42. Reznick R, Regehr G, MacRae H, Martin J, McCulloch W. Testing technical skill
via an innovative "bench station" examination. Am J Surg 1997;173(3):226-30.
43. MacRae H, Regehr G, Leadbetter W, Reznick RK. A comprehensive examination
for senior surgical residents. Am J Surg 2000;179(3):190-3.
44. Regehr G, MacRae H, Reznick RK, Szalay D. Comparing the psychometric
properties of checklists and global rating scales for assessing performance on an
OSCE-format examination. Acad Med 1998;73(9):993-7.
45. Lehmann KS, Ritz JP, Maass H, et al. A prospective randomized study to test the
transfer of basic psychomotor skills from virtual reality to physical reality in a
comparable training setting. Ann Surg 2005;241(3):442-9.
46. Rotnes JS, Kaasa J, Westgaard G, et al. A tutorial platform suitable for surgical
simulator training (SimMentor). Stud Health Technol Inform 2002;85:419-25.
47. Strom P, Kjellin A, Hedman L, Wredmark T, Fellander-Tsai L. Training in tasks
with different visual-spatial components does not improve virtual arthroscopy
performance. Surg Endosc 2004;18(1):115-20.
48. Di Lorenzo N DJ. Surgical Training and Simulations. Business briefing: Global
Surgery--Future Directions 2005.
49. Gallagher AG, Ritter EM, Satava RM. Fundamental principles of validation, and
reliability: rigorous science for the assessment of surgical education and training.
Surg Endosc 2003;17(10):1525-9.
50. Gould DA, Reekers JA, Kessel DO, et al. Simulation devices in interventional
radiology: caveat emptor. Cardiovasc Intervent Radiol 2006;29(1):4-6.
51. Torkington J, Smith SG, Rees BI, Darzi A. Skill transfer from virtual reality to a
real laparoscopic task. Surg Endosc 2001;15(10):1076-9.
76
52. Ahlberg G, Heikkinen T, Iselius L, Leijonmarck CE, Rutqvist J, Arvidsson D. Does
training in a virtual reality simulator improve surgical performance? Surg
Endosc 2002;16(1):126-9.
53. Schijven MP, Jakimowicz JJ, Broeders IA, Tseng LN. The Eindhoven laparoscopic
cholecystectomy training course--improving operating room performance using
virtual reality training: results from the first E.A.E.S. accredited virtual reality
trainings curriculum. Surg Endosc 2005;19(9):1220-6.
54. Representatives USSaUSHo. Patient and Physician Safety and Protection Act of
2005. S.1297 / H.R.1228, 2005.
55. ACGME ACoGME. Common Program Requirements, 2003.
56. AAMC AoAMC, ed. AAMC Policy Guidance on Graduate Medical Education:
Assuring Quality Patient Care and Quality Education, 2001.
57. AMA AMA. H-310.979 Resident Physician Working Hours and Supervision. 2005.
58. EU TCotEU. Working Time Directive, 1993.
59. Russell WMS, Burch RL. The principles of humane experimental technique.
London: Methuen, 1959.
60. Orlans FB. Ethical decision making about animal experiments. Ethics Behav
1997;7(2):163-71.
61. Bernstein MH. Without a tear : our tragic relationship with animals. Urbana:
University of Illinois Press, 2004.
62. DeGrazia D. The ethics of animal research: what are the prospects for agreement?
Camb Q Healthc Ethics 1999;8(1):23-34.
63. Regan T. The case for animal rights. Updated with a new preface ed. Berkeley:
University of California Press, 2004.
64. Grantcharov TP, Bardram L, Funch-Jensen P, Rosenberg J. Learning curves and
impact of previous operative experience on performance on a virtual reality
simulator to test laparoscopic surgical skills. Am J Surg 2003;185(2):146-9.
65. Seymour NE, Gallagher AG, Roman SA, et al. Virtual reality training improves
operating room performance: results of a randomized, double-blinded study.
Ann Surg 2002;236(4):458-63; discussion 463-4.
66. Strom P, Kjellin A, Hedman L, Johnson E, Wredmark T, Fellander-Tsai L.
Validation and learning in the Procedicus KSA virtual reality surgical simulator.
Surg Endosc 2003;17(2):227-31.
67. Gallagher AG, Smith CD, Bowers SP, et al. Psychomotor skills assessment in
practicing surgeons experienced in performing advanced laparoscopic
procedures. J Am Coll Surg 2003;197(3):479-88.
77
68. NIH. Cost-Benefit Analysis Guide for NIH IT Projects. In: Services HaH, ed:
Deputy Chief Information Officer, Center for Information Technology, 1999.
69. R: A Language and Environment for Statistical Computing [program]. Vienna,
Austria, 2005.
70. Zeger SL, Liang, K.Y. Longitudinal data analysis for discrete and continuous
outcomes. Biometrics 1986(42):121-30.
71. Grantcharov TP, Rosenberg J, Pahle E, Funch-Jensen P. Virtual reality computer
simulation. Surg Endosc 2001;15(3):242-4.
72. National Institutes of Health (U.S.). NIH Guide for Grants and Contracts, Vol. 14,
No. 8, 25 Jun 1985. Special Edition : Laboratory Animal Welfare. Bethesda,
Md.: National Institutes of Health, 1985.
73. Grantcharov TP, Bardram L, Funch-Jensen P, Rosenberg J. Impact of hand
dominance, gender, and experience with computer games on performance in
virtual reality laparoscopy. Surg Endosc 2003;17(7):1082-5.
74. McClusky DA, 3rd, Ritter EM, Lederman AB, Gallagher AG, Smith CD.
Correlation between perceptual, visuo-spatial, and psychomotor aptitude to
duration of training required to reach performance goals on the MIST-VR
surgical simulator. Am Surg 2005;71(1):13-20; discussion 20-1.
75. Arora H, Uribe J, Ralph W, et al. Assessment of construct validity of the endoscopic
sinus surgery simulator. Arch Otolaryngol Head Neck Surg 2005;131(3):217-21.
76. Enochsson L, Westman B, Ritter EM, et al. Objective assessment of visuospatial and
psychomotor ability and flow of residents and senior endoscopists in simulated
gastroscopy. Surg Endosc 2006.
77. Dayal R, Faries PL, Lin SC, et al. Computer simulation as a component of catheter-
based training. J Vasc Surg 2004;40(6):1112-7.
78. Hsu JH, Younan D, Pandalai S, et al. Use of computer simulation for determining
endovascular skill levels in a carotid stenting model. J Vasc Surg
2004;40(6):1118-25.
79. Blyth P, Anderson IA, Stott NS. Virtual reality simulators in orthopedic surgery:
what do the surgeons think? J Surg Res 2006;131(1):133-9.
80. Aggarwal R, Black SA, Hance JR, Darzi A, Cheshire NJ. Virtual reality simulation
training can improve inexperienced surgeons' endovascular skills. Eur J Vasc
Endovasc Surg 2006;31(6):588-93.
78
81. Gallagher AG, Cates CU. Approval of virtual reality training for carotid stenting:
what this means for procedural-based medicine. Jama 2004;292(24):3024-6.
82. Berry M, Lystig T, Reznick R, Lonn L. Assessment of a virtual interventional
simulator trainer. J Endovasc Ther 2006;13(2):237-43.
83. Chaer RA, Derubertis BG, Lin SC, et al. Simulation improves resident performance
in catheter-based intervention: results of a randomized, controlled study. Ann
Surg 2006;244(3):343-52.
84. Johnson S, Healey A, Evans J, Murphy M, Crawshaw M, Gould D. Physical and
cognitive task analysis in interventional radiology. Clin Radiol 2006;61(1):97-103.
85. Gould DA, Reekers JA, Kessel DO, et al. Simulation devices in interventional
radiology: validation pending. J Vasc Interv Radiol 2006;17(2 Pt 1):215-6.
86. Gould DA, Healey AE, Johnson SJ, Lewandowski WE, Kessel DO. Metrics for an
interventional radiology curriculum: a case for standardisation? Stud Health
Technol Inform 2006;119:159-64.
87. Dawson DL. Training in carotid artery stenting: do carotid simulation systems
really help? Vascular 2006;14(5):256-63.
88. Hamilton EC, Scott DJ, Fleming JB, et al. Comparison of video trainer and virtual
reality training systems on acquisition of laparoscopic skills. Surg Endosc
2002;16(3):406-11.
89. Szalay D, MacRae H, Regehr G, Reznick R. Using operative outcome to assess
technical skill. Am J Surg 2000;180(3):234-7.
90. Maithel S, Sierra R, Korndorffer J, et al. Construct and face validity of MIST-VR,
Endotower, and CELTS: are we ready for skills assessment using simulators?
Surg Endosc 2006;20(1):104-12.
91. Grober ED, Hamstra SJ, Wanzel KR, et al. Validation of novel and objective
measures of microsurgical skill: Hand-motion analysis and stereoscopic visual
acuity. Microsurgery 2003;23(4):317-22.
92. Aggarwal R, Grantcharov T, Moorthy K, Hance J, Darzi A. A competency-based
virtual reality training curriculum for the acquisition of laparoscopic
psychomotor skill. Am J Surg 2006;191(1):128-133.
93. Wang EE, Vozenilek JA. Addressing the systems-based practice core competency: a
simulation-based curriculum. Acad Emerg Med 2005;12(12):1191-4.
79
94. Madan AK, Frantzides CT, Quiros R, Dujovny N, Tebbit C. Effects of a
laparoscopic course on student interest in surgical residency. Jsls 2005;9(2):134-
7.
95. Beard J, Strachan A, Davies H, et al. Developing an education and assessment
framework for the Foundation Programme. Med Educ 2005;39(8):841-51.
96. Berry M, Lystig T, Beard J, Klingenstierna H, Reznick R, Lonn L. Porcine Transfer
Study: Virtual Reality Simulator Training Compared with Porcine Training in
Endovascular Novices. Cardiovasc Intervent Radiol 2007;30(3):In press.
97. Ziv A, Wolpe PR, Small SD, Glick S. Simulation-based medical education: an ethical
imperative. Acad Med 2003;78(8):783-8.
98. Gruber FP, Hartung T. Alternatives to animal experimentation in basic research.
Altex 2004;21 Suppl 1:3-31.
99. Fitzpatrick A. Ethics and animal research. J Lab Clin Med 2003;141(2):89-90.
100. Machan TR. Putting humans first : why we are nature's favorite. Lanham, Md.:
Rowman & Littlefield Publishers, 2004.
101. Schijven M, Jakimowicz J. Virtual reality surgical laparoscopic simulators. Surg
Endosc 2003;17(12):1943-50.
80